# The LearningOnline Network with CAPA # # $Id: loncoursedata.pm,v 1.58 2003/03/20 19:27:26 matthew Exp $ # # Copyright Michigan State University Board of Trustees # # This file is part of the LearningOnline Network with CAPA (LON-CAPA). # # LON-CAPA is free software; you can redistribute it and/or modify # it under the terms of the GNU General Public License as published by # the Free Software Foundation; either version 2 of the License, or # (at your option) any later version. # # LON-CAPA is distributed in the hope that it will be useful, # but WITHOUT ANY WARRANTY; without even the implied warranty of # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the # GNU General Public License for more details. # # You should have received a copy of the GNU General Public License # along with LON-CAPA; if not, write to the Free Software # Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA # # /home/httpd/html/adm/gpl.txt # # http://www.lon-capa.org/ # ### =pod =head1 NAME loncoursedata =head1 SYNOPSIS Set of functions that download and process student and course information. =head1 PACKAGES USED Apache::Constants qw(:common :http) Apache::lonnet() Apache::lonhtmlcommon HTML::TokeParser GDBM_File =cut package Apache::loncoursedata; use strict; use Apache::Constants qw(:common :http); use Apache::lonnet(); use Apache::lonhtmlcommon; use Time::HiRes; use Apache::lonmysql; use HTML::TokeParser; use GDBM_File; =pod =head1 DOWNLOAD INFORMATION This section contains all the functions that get data from other servers and/or itself. =cut # ----- DOWNLOAD INFORMATION ------------------------------------------- =pod =item &DownloadClasslist() Collects lastname, generation, middlename, firstname, PID, and section for each student from their environment database. The section data is also download, though it is in a rough format, and is processed later. The list of students is built from collecting a classlist for the course that is to be displayed. Once the classlist has been downloaded, its date stamp is recorded. Unless the datestamp for the class database is reset or is modified, this data will not be downloaded again. Also, there was talk about putting the fullname and section and perhaps other pieces of data into the classlist file. This would reduce the number of different file accesses and reduce the amount of processing on this side. =over 4 Input: $courseID, $lastDownloadTime, $c $courseID: The id of the course $lastDownloadTime: This is the date stamp for when this information was last gathered. If it is set to Not downloaded, it will gather the data again, though it currently does not remove the old data. $c: The connection class that can determine if the browser has aborted. It is used to short circuit this function so that it does not continue to get information when there is no need. Output: \%classlist \%classlist: A pointer to a hash containing the following data: -A list of student name:domain (as keys) (known below as $name) -A hash pointer for each student containing lastname, generation, firstname, middlename, and PID : Key is $name.studentInformation -A hash pointer to each students section data : Key is $name.section -If there was an error in dump, it will be returned in the hash. See the error codes for dump in lonnet. Also, an error key will be generated if an abort occurs. =back =cut sub DownloadClasslist { my ($courseID, $lastDownloadTime, $c)=@_; my ($courseDomain,$courseNumber)=split(/\_/,$courseID); my %classlist; my $modifiedTime = &Apache::lonnet::GetFileTimestamp($courseDomain, $courseNumber, 'classlist.db', $Apache::lonnet::perlvar{'lonUsersDir'}); # Always download the information if lastDownloadTime is set to # Not downloaded, otherwise it is only downloaded if the file # has been updated and has a more recent date stamp if($lastDownloadTime ne 'Not downloaded' && $lastDownloadTime >= $modifiedTime && $modifiedTime >= 0) { # Data is not gathered so return UpToDate as true. This # will be interpreted in ProcessClasslist $classlist{'lastDownloadTime'}=time; $classlist{'UpToDate'} = 'true'; return \%classlist; } %classlist=&Apache::lonnet::dump('classlist',$courseDomain, $courseNumber); foreach(keys (%classlist)) { if(/^(con_lost|error|no_such_host)/i) { return; } } foreach my $name (keys(%classlist)) { if(defined($c) && ($c->aborted())) { $classlist{'error'}='aborted'; return \%classlist; } my ($studentName,$studentDomain) = split(/\:/,$name); # Download student environment data, specifically the full name and id. my %studentInformation=&Apache::lonnet::get('environment', ['lastname','generation', 'firstname','middlename', 'id'], $studentDomain, $studentName); $classlist{$name.':studentInformation'}=\%studentInformation; if($c->aborted()) { $classlist{'error'}='aborted'; return \%classlist; } #Section my %section=&Apache::lonnet::dump('roles',$studentDomain,$studentName); $classlist{$name.':sections'}=\%section; } $classlist{'UpToDate'} = 'false'; $classlist{'lastDownloadTime'}=time; return \%classlist; } =pod =item &DownloadCourseInformation() Dump of all the course information for a single student. The data can be pruned by making use of dumps regular expression arguement. This function also takes a regular expression which it passes straight through to dump. The data is no escaped, because it is done elsewhere. It also checks the timestamp of the students course database file and only downloads if it has been modified since the last download. =over 4 Input: $namedata, $courseID, $lastDownloadTime, $WhatIWant $namedata: student name:domain $courseID: The id of the course $lastDownloadTime: This is the date stamp for when this information was last gathered. If it is set to Not downloaded, it will gather the data again, though it currently does not remove the old data. $WhatIWant: Regular expression used to get selected data with dump Output: \%courseData \%courseData: A hash pointer to the raw data from the students course database. =back =cut sub DownloadCourseInformation { my ($namedata,$courseID,$lastDownloadTime,$WhatIWant)=@_; my %courseData; my ($name,$domain) = split(/\:/,$namedata); my $modifiedTime = &Apache::lonnet::GetFileTimestamp($domain, $name, $courseID.'.db', $Apache::lonnet::perlvar{'lonUsersDir'}); if($lastDownloadTime ne 'Not downloaded' && $lastDownloadTime >= $modifiedTime && $modifiedTime >= 0) { # Data is not gathered so return UpToDate as true. This # will be interpreted in ProcessClasslist $courseData{$namedata.':lastDownloadTime'}=time; $courseData{$namedata.':UpToDate'} = 'true'; return \%courseData; } # Download course data if(!defined($WhatIWant)) { # set the regular expression to everything by setting it to period $WhatIWant = '.'; } %courseData=&Apache::lonnet::dump($courseID, $domain, $name, $WhatIWant); $courseData{'UpToDate'} = 'false'; $courseData{'lastDownloadTime'}=time; my %newData; foreach (keys(%courseData)) { # need to have the keys to be prepended with the name:domain of the # student to reduce data collision later. $newData{$namedata.':'.$_} = $courseData{$_}; } return \%newData; } # ----- END DOWNLOAD INFORMATION --------------------------------------- =pod =head1 PROCESSING FUNCTIONS These functions process all the data for all the students. Also, they are the functions that access the cache database for writing the majority of the time. The downloading and caching were separated to reduce problems with stopping downloading then can not tie hash to database later. =cut # ----- PROCESSING FUNCTIONS --------------------------------------- #################################################### #################################################### =pod =item &get_sequence_assessment_data() AT THIS TIME THE USE OF THIS FUNCTION IS *NOT* RECOMMENDED Use lonnavmaps to build a data structure describing the order and assessment contents of each sequence in the current course. The returned structure is a hash reference. { title => 'title', symb => 'symb', source => '/s/o/u/r/c/e', type => (container|assessment), num_assess => 2, # only for container parts => [11,13,15], # only for assessment response_ids => [12,14,16], # only for assessment contents => [........] # only for container } $hash->{'contents'} is a reference to an array of hashes of the same structure. Also returned are array references to the sequences and assessments contained in the course. =cut #################################################### #################################################### sub get_sequence_assessment_data { my $fn=$ENV{'request.course.fn'}; ## ## use navmaps my $navmap = Apache::lonnavmaps::navmap->new($fn.".db",$fn."_parms.db", 1,0); if (!defined($navmap)) { return 'Can not open Coursemap'; } my $iterator = $navmap->getIterator(undef, undef, undef, 1); ## ## Prime the pump ## ## We are going to loop until we run out of sequences/pages to explore for ## resources. This means we have to start out with something to look ## at. my $title = $ENV{'course.'.$ENV{'request.course.id'}.'.description'}; my $symb = 'top'; my $src = 'not applicable'; # my @Sequences; my @Assessments; my @Nested_Sequences = (); # Stack of sequences, keeps track of depth my $top = { title => $title, src => $src, symb => $symb, type => 'container', num_assess => 0, num_assess_parts => 0, contents => [], }; push (@Sequences,$top); push (@Nested_Sequences, $top); # # We need to keep track of which sequences contain homework problems # my $previous; my $curRes = $iterator->next(); # BEGIN_MAP $curRes = $iterator->next(); # The first item in the top level map. while (scalar(@Nested_Sequences)) { $previous = $curRes; $curRes = $iterator->next(); my $currentmap = $Nested_Sequences[-1]; # Last one on the stack if ($curRes == $iterator->BEGIN_MAP()) { # get the map itself, instead of BEGIN_MAP $title = $previous->title(); $symb = $previous->symb(); $src = $previous->src(); my $newmap = { title => $title, src => $src, symb => $symb, type => 'container', num_assess => 0, contents => [], }; push (@{$currentmap->{'contents'}},$newmap); # this is permanent push (@Sequences,$newmap); push (@Nested_Sequences, $newmap); # this is a stack next; } if ($curRes == $iterator->END_MAP()) { pop(@Nested_Sequences); next; } next if (! ref($curRes)); next if (! $curRes->is_problem());# && !$curRes->randomout); # Okay, from here on out we only deal with assessments $title = $curRes->title(); $symb = $curRes->symb(); $src = $curRes->src(); my $parts = $curRes->parts(); my $assessment = { title => $title, src => $src, symb => $symb, type => 'assessment', parts => $parts, num_parts => scalar(@$parts), }; push(@Assessments,$assessment); push(@{$currentmap->{'contents'}},$assessment); $currentmap->{'num_assess'}++; $currentmap->{'num_assess_parts'}+= scalar(@$parts); } $navmap->untieHashes(); return ($top,\@Sequences,\@Assessments); } ################################################# ################################################# =pod =item &ProcessTopResourceMap() Trace through the "big hash" created in rat/lonuserstate.pm::loadmap. Basically, this function organizes a subset of the data and stores it in cached data. The data stored is the problems, sequences, sequence titles, parts of problems, and their ordering. Column width information is also partially handled here on a per sequence basis. =over 4 Input: $cache, $c $cache: A pointer to a hash to store the information $c: The connection class used to determine if an abort has been sent to the browser Output: A string that contains an error message or "OK" if everything went smoothly. =back =cut sub ProcessTopResourceMap { my ($cache,$c)=@_; my %hash; my $fn=$ENV{'request.course.fn'}; if(-e "$fn.db") { my $tieTries=0; while($tieTries < 3) { if($c->aborted()) { return; } if(tie(%hash,'GDBM_File',"$fn.db",&GDBM_READER(),0640)) { last; } $tieTries++; sleep 1; } if($tieTries >= 3) { return 'Coursemap undefined.'; } } else { return 'Can not open Coursemap.'; } my $oldkeys; delete $cache->{'OptionResponses'}; if(defined($cache->{'ResourceKeys'})) { $oldkeys = $cache->{'ResourceKeys'}; foreach (split(':::', $cache->{'ResourceKeys'})) { delete $cache->{$_}; } delete $cache->{'ResourceKeys'}; } # Initialize state machine. Set information pointing to top level map. my (@sequences, @currentResource, @finishResource); my ($currentSequence, $currentResourceID, $lastResourceID); $currentResourceID=$hash{'ids_'. &Apache::lonnet::clutter($ENV{'request.course.uri'})}; push(@currentResource, $currentResourceID); $lastResourceID=-1; $currentSequence=-1; my $topLevelSequenceNumber = $currentSequence; my %sequenceRecord; my %allkeys; while(1) { if($c->aborted()) { last; } # HANDLE NEW SEQUENCE! #if page || sequence if(defined($hash{'map_pc_'.$hash{'src_'.$currentResourceID}}) && !defined($sequenceRecord{$currentResourceID})) { $sequenceRecord{$currentResourceID}++; push(@sequences, $currentSequence); push(@currentResource, $currentResourceID); push(@finishResource, $lastResourceID); $currentSequence=$hash{'map_pc_'.$hash{'src_'.$currentResourceID}}; # Mark sequence as containing problems. If it doesn't, then # it will be removed when processing for this sequence is # complete. This allows the problems in a sequence # to be outputed before problems in the subsequences if(!defined($cache->{'orderedSequences'})) { $cache->{'orderedSequences'}=$currentSequence; } else { $cache->{'orderedSequences'}.=':'.$currentSequence; } $allkeys{'orderedSequences'}++; $lastResourceID=$hash{'map_finish_'. $hash{'src_'.$currentResourceID}}; $currentResourceID=$hash{'map_start_'. $hash{'src_'.$currentResourceID}}; if(!($currentResourceID) || !($lastResourceID)) { $currentSequence=pop(@sequences); $currentResourceID=pop(@currentResource); $lastResourceID=pop(@finishResource); if($currentSequence eq $topLevelSequenceNumber) { last; } } next; } # Handle gradable resources: exams, problems, etc $currentResourceID=~/(\d+)\.(\d+)/; my $partA=$1; my $partB=$2; if($hash{'src_'.$currentResourceID}=~ /\.(problem|exam|quiz|assess|survey|form)$/ && $partA eq $currentSequence && !defined($sequenceRecord{$currentSequence.':'. $currentResourceID})) { $sequenceRecord{$currentSequence.':'.$currentResourceID}++; my $Problem = &Apache::lonnet::symbclean( &Apache::lonnet::declutter($hash{'map_id_'.$partA}). '___'.$partB.'___'. &Apache::lonnet::declutter($hash{'src_'. $currentResourceID})); $cache->{$currentResourceID.':problem'}=$Problem; $allkeys{$currentResourceID.':problem'}++; if(!defined($cache->{$currentSequence.':problems'})) { $cache->{$currentSequence.':problems'}=$currentResourceID; } else { $cache->{$currentSequence.':problems'}.= ':'.$currentResourceID; } $allkeys{$currentSequence.':problems'}++; my $meta=$hash{'src_'.$currentResourceID}; # $cache->{$currentResourceID.':title'}= # &Apache::lonnet::metdata($meta,'title'); $cache->{$currentResourceID.':title'}= $hash{'title_'.$currentResourceID}; $allkeys{$currentResourceID.':title'}++; $cache->{$currentResourceID.':source'}= $hash{'src_'.$currentResourceID}; $allkeys{$currentResourceID.':source'}++; # Get Parts for problem my %beenHere; foreach (split(/\,/,&Apache::lonnet::metadata($meta,'packages'))) { if(/^\w+response_\d+.*/) { my (undef, $partId, $responseId) = split(/_/,$_); if($beenHere{'p:'.$partId} == 0) { $beenHere{'p:'.$partId}++; if(!defined($cache->{$currentSequence.':'. $currentResourceID.':parts'})) { $cache->{$currentSequence.':'.$currentResourceID. ':parts'}=$partId; } else { $cache->{$currentSequence.':'.$currentResourceID. ':parts'}.=':'.$partId; } $allkeys{$currentSequence.':'.$currentResourceID. ':parts'}++; } if($beenHere{'r:'.$partId.':'.$responseId} == 0) { $beenHere{'r:'.$partId.':'.$responseId}++; if(!defined($cache->{$currentSequence.':'. $currentResourceID.':'.$partId. ':responseIDs'})) { $cache->{$currentSequence.':'.$currentResourceID. ':'.$partId.':responseIDs'}=$responseId; } else { $cache->{$currentSequence.':'.$currentResourceID. ':'.$partId.':responseIDs'}.=':'. $responseId; } $allkeys{$currentSequence.':'.$currentResourceID.':'. $partId.':responseIDs'}++; } if(/^optionresponse/ && $beenHere{'o:'.$partId.':'.$currentResourceID} == 0) { $beenHere{'o:'.$partId.$currentResourceID}++; if(defined($cache->{'OptionResponses'})) { $cache->{'OptionResponses'}.= ':::'. $currentSequence.':'.$currentResourceID.':'. $partId.':'.$responseId; } else { $cache->{'OptionResponses'}= $currentSequence.':'. $currentResourceID.':'. $partId.':'.$responseId; } $allkeys{'OptionResponses'}++; } } } } # if resource == finish resource, then it is the end of a sequence/page if($currentResourceID eq $lastResourceID) { # pop off last resource of sequence $currentResourceID=pop(@currentResource); $lastResourceID=pop(@finishResource); if(defined($cache->{$currentSequence.':problems'})) { # Capture sequence information here $cache->{$currentSequence.':title'}= $hash{'title_'.$currentResourceID}; $allkeys{$currentSequence.':title'}++; $cache->{$currentSequence.':source'}= $hash{'src_'.$currentResourceID}; $allkeys{$currentSequence.':source'}++; my $totalProblems=0; foreach my $currentProblem (split(/\:/, $cache->{$currentSequence. ':problems'})) { foreach (split(/\:/,$cache->{$currentSequence.':'. $currentProblem. ':parts'})) { $totalProblems++; } } my @titleLength=split(//,$cache->{$currentSequence. ':title'}); # $extra is 5 for problems correct and 3 for space # between problems correct and problem output my $extra = 8; if(($totalProblems + $extra) > (scalar @titleLength)) { $cache->{$currentSequence.':columnWidth'}= $totalProblems + $extra; } else { $cache->{$currentSequence.':columnWidth'}= (scalar @titleLength); } $allkeys{$currentSequence.':columnWidth'}++; } else { # Remove sequence from list, if it contains no problems to # display. $cache->{'orderedSequences'}=~s/$currentSequence//; $cache->{'orderedSequences'}=~s/::/:/g; $cache->{'orderedSequences'}=~s/^:|:$//g; } $currentSequence=pop(@sequences); if($currentSequence eq $topLevelSequenceNumber) { last; } } # MOVE!!! # move to next resource unless(defined($hash{'to_'.$currentResourceID})) { # big problem, need to handle. Next is probably wrong my $errorMessage = 'Big problem in '; $errorMessage .= 'loncoursedata::ProcessTopLevelMap.'; $errorMessage .= " bighash to_$currentResourceID not defined!"; &Apache::lonnet::logthis($errorMessage); if (!defined($currentResourceID)) {last;} } my @nextResources=(); foreach (split(/\,/,$hash{'to_'.$currentResourceID})) { if(!defined($sequenceRecord{$currentSequence.':'. $hash{'goesto_'.$_}})) { push(@nextResources, $hash{'goesto_'.$_}); } } push(@currentResource, @nextResources); # Set the next resource to be processed $currentResourceID=pop(@currentResource); } my @theKeys = keys(%allkeys); my $newkeys = join(':::', @theKeys); $cache->{'ResourceKeys'} = join(':::', $newkeys); if($newkeys ne $oldkeys) { $cache->{'ResourceUpdated'} = 'true'; } else { $cache->{'ResourceUpdated'} = 'false'; } unless (untie(%hash)) { &Apache::lonnet::logthis("WARNING: ". "Could not untie coursemap $fn (browse)". "."); } return 'OK'; } =pod =item &ProcessClasslist() Taking the class list dumped from &DownloadClasslist(), all the students and their non-class information is processed using the &ProcessStudentInformation() function. A date stamp is also recorded for when the data was processed. Takes data downloaded for a student and breaks it up into managable pieces and stored in cache data. The username, domain, class related date, PID, full name, and section are all processed here. =over 4 Input: $cache, $classlist, $courseID, $ChartDB, $c $cache: A hash pointer to store the data $classlist: The hash of data collected about a student from &DownloadClasslist(). The hash contains a list of students, a pointer to a hash of student information for each student, and each students section number. $courseID: The course ID $ChartDB: The name of the cache database file. $c: The connection class used to determine if an abort has been sent to the browser Output: @names @names: An array of students whose information has been processed, and are to be considered in an arbitrary order. The entries in @names are of the form username:domain. The values in $cache are as follows: *NOTE: for the following $name implies username:domain $name.':error' only defined if an error occured. Value contains the error message $name.':lastDownloadTime' unconverted time of the last update of a student\'s course data $name.'updateTime' coverted time of the last update of a student\'s course data $name.':username' username of a student $name.':domain' domain of a student $name.':fullname' full name of a student $name.':id' PID of a student $name.':Status' active/expired status of a student $name.':section' section of a student =back =cut sub ProcessClasslist { my ($cache,$classlist,$courseID,$c)=@_; my @names=(); $cache->{'ClasslistTimeStamp'}=$classlist->{'lastDownloadTime'}; if($classlist->{'UpToDate'} eq 'true') { return split(/:::/,$cache->{'NamesOfStudents'});; } foreach my $name (keys(%$classlist)) { if($name =~ /\:section/ || $name =~ /\:studentInformation/ || $name eq '' || $name eq 'UpToDate' || $name eq 'lastDownloadTime') { next; } if($c->aborted()) { return (); } my $studentInformation = $classlist->{$name.':studentInformation'}; my $date = $classlist->{$name}; my ($studentName,$studentDomain) = split(/\:/,$name); $cache->{$name.':username'}=$studentName; $cache->{$name.':domain'}=$studentDomain; # Initialize timestamp for student if(!defined($cache->{$name.':lastDownloadTime'})) { $cache->{$name.':lastDownloadTime'}='Not downloaded'; $cache->{$name.':updateTime'}=' Not updated'; } my $error = 0; foreach(keys(%$studentInformation)) { if(/^(con_lost|error|no_such_host)/i) { $cache->{$name.':error'}= 'Could not download student environment data.'; $cache->{$name.':fullname'}=''; $cache->{$name.':id'}=''; $error = 1; } } next if($error); push(@names,$name); $cache->{$name.':fullname'}=&ProcessFullName( $studentInformation->{'lastname'}, $studentInformation->{'generation'}, $studentInformation->{'firstname'}, $studentInformation->{'middlename'}); $cache->{$name.':id'}=$studentInformation->{'id'}; my ($end, $start)=split(':',$date); $courseID=~s/\_/\//g; $courseID=~s/^(\w)/\/$1/; my $sec=''; my $sectionData = $classlist->{$name.':sections'}; foreach my $key (keys (%$sectionData)) { my $value = $sectionData->{$key}; if ($key=~/^$courseID(?:\/)*(\w+)*\_st$/) { my $tempsection=$1; if($key eq $courseID.'_st') { $tempsection=''; } my (undef,$roleend,$rolestart)=split(/\_/,$value); if($roleend eq $end && $rolestart eq $start) { $sec = $tempsection; last; } } } my $status='Expired'; if(((!$end) || time < $end) && ((!$start) || (time > $start))) { $status='Active'; } $cache->{$name.':Status'}=$status; $cache->{$name.':section'}=$sec; if($sec eq '' || !defined($sec) || $sec eq ' ') { $sec = 'none'; } if(defined($cache->{'sectionList'})) { if($cache->{'sectionList'} !~ /(^$sec:|^$sec$|:$sec$|:$sec:)/) { $cache->{'sectionList'} .= ':'.$sec; } } else { $cache->{'sectionList'} = $sec; } } $cache->{'ClasslistTimestamp'}=time; $cache->{'NamesOfStudents'}=join(':::',@names); return @names; } =pod =item &ProcessStudentData() Takes the course data downloaded for a student in &DownloadCourseInformation() and breaks it up into key value pairs to be stored in the cached data. The keys are comprised of the $username:$domain:$keyFromCourseDatabase. The student username:domain is stored away signifying that the students information has been downloaded and can be reused from cached data. =over 4 Input: $cache, $courseData, $name $cache: A hash pointer to store data $courseData: A hash pointer that points to the course data downloaded for a student. $name: username:domain Output: None *NOTE: There is no output, but an error message is stored away in the cache data. This is checked in &FormatStudentData(). The key username:domain:error will only exist if an error occured. The error is an error from &DownloadCourseInformation(). =back =cut sub ProcessStudentData { my ($cache,$courseData,$name)=@_; if(!&CheckDateStampError($courseData, $cache, $name)) { return; } # This little delete thing, should not be here. Move some other # time though. if(defined($cache->{$name.':keys'})) { foreach (split(':::', $cache->{$name.':keys'})) { delete $cache->{$name.':'.$_}; } delete $cache->{$name.':keys'}; } my %courseKeys; # user name:domain was prepended earlier in DownloadCourseInformation foreach (keys %$courseData) { my $currentKey = $_; $currentKey =~ s/^$name//; $courseKeys{$currentKey}++; $cache->{$_}=$courseData->{$_}; } $cache->{$name.':keys'} = join(':::', keys(%courseKeys)); return; } =pod =item &ExtractStudentData() HISTORY: This function originally existed in every statistics module, and performed different tasks, the had some overlap. Due to the need for the data from the different modules, they were combined into a single function. This function now extracts all the necessary course data for a student from what was downloaded from their homeserver. There is some extra time overhead compared to the ProcessStudentInformation function, but it would have had to occurred at some point anyways. This is now typically called while downloading the data it will process. It is the brother function to ProcessStudentInformation. =over 4 Input: $input, $output, $data, $name $input: A hash that contains the input data to be processed $output: A hash to contain the processed data $data: A hash containing the information on what is to be processed and how (basically). $name: username:domain The input is slightly different here, but is quite simple. It is currently used where the $input, $output, and $data can and are often the same hashes, but they do not need to be. Output: None *NOTE: There is no output, but an error message is stored away in the cache data. This is checked in &FormatStudentData(). The key username:domain:error will only exist if an error occured. The error is an error from &DownloadCourseInformation(). =back =cut sub ExtractStudentData { my ($input, $output, $data, $name)=@_; if(!&CheckDateStampError($input, $data, $name)) { return; } # This little delete thing, should not be here. Move some other # time though. my %allkeys; if(defined($output->{$name.':keys'})) { foreach (split(':::', $output->{$name.':keys'})) { delete $output->{$name.':'.$_}; } delete $output->{$name.':keys'}; } my ($username,$domain)=split(':',$name); my $Version; my $problemsCorrect = 0; my $totalProblems = 0; my $problemsSolved = 0; my $numberOfParts = 0; my $totalAwarded = 0; foreach my $sequence (split(':', $data->{'orderedSequences'})) { foreach my $problemID (split(':', $data->{$sequence.':problems'})) { my $problem = $data->{$problemID.':problem'}; my $LatestVersion = $input->{$name.':version:'.$problem}; # Output dashes for all the parts of this problem if there # is no version information about the current problem. $output->{$name.':'.$problemID.':NoVersion'} = 'false'; $allkeys{$name.':'.$problemID.':NoVersion'}++; if(!$LatestVersion) { foreach my $part (split(/\:/,$data->{$sequence.':'. $problemID. ':parts'})) { $output->{$name.':'.$problemID.':'.$part.':tries'} = 0; $output->{$name.':'.$problemID.':'.$part.':awarded'} = 0; $output->{$name.':'.$problemID.':'.$part.':code'} = ' '; $allkeys{$name.':'.$problemID.':'.$part.':tries'}++; $allkeys{$name.':'.$problemID.':'.$part.':awarded'}++; $allkeys{$name.':'.$problemID.':'.$part.':code'}++; $totalProblems++; } $output->{$name.':'.$problemID.':NoVersion'} = 'true'; next; } my %partData=undef; # Initialize part data, display skips correctly # Skip refers to when a student made no submissions on that # part/problem. foreach my $part (split(/\:/,$data->{$sequence.':'. $problemID. ':parts'})) { $partData{$part.':tries'}=0; $partData{$part.':code'}=' '; $partData{$part.':awarded'}=0; $partData{$part.':timestamp'}=0; foreach my $response (split(':', $data->{$sequence.':'. $problemID.':'. $part.':responseIDs'})) { $partData{$part.':'.$response.':submission'}=''; } } # Looping through all the versions of each part, starting with the # oldest version. Basically, it gets the most recent # set of grade data for each part. my @submissions = (); for(my $Version=1; $Version<=$LatestVersion; $Version++) { foreach my $part (split(/\:/,$data->{$sequence.':'. $problemID. ':parts'})) { if(!defined($input->{"$name:$Version:$problem". ":resource.$part.solved"})) { # No grade for this submission, so skip next; } my $tries=0; my $code=' '; my $awarded=0; $tries = $input->{$name.':'.$Version.':'.$problem. ':resource.'.$part.'.tries'}; $awarded = $input->{$name.':'.$Version.':'.$problem. ':resource.'.$part.'.awarded'}; $partData{$part.':awarded'}=($awarded) ? $awarded : 0; $partData{$part.':tries'}=($tries) ? $tries : 0; $partData{$part.':timestamp'}=$input->{$name.':'.$Version.':'. $problem. ':timestamp'}; if(!$input->{$name.':'.$Version.':'.$problem.':resource.'.$part. '.previous'}) { foreach my $response (split(':', $data->{$sequence.':'. $problemID.':'. $part.':responseIDs'})) { @submissions=($input->{$name.':'.$Version.':'. $problem. ':resource.'.$part.'.'. $response.'.submission'}, @submissions); } } my $val = $input->{$name.':'.$Version.':'.$problem. ':resource.'.$part.'.solved'}; if ($val eq 'correct_by_student') {$code = '*';} elsif ($val eq 'correct_by_override') {$code = '+';} elsif ($val eq 'incorrect_attempted') {$code = '.';} elsif ($val eq 'incorrect_by_override'){$code = '-';} elsif ($val eq 'excused') {$code = 'x';} elsif ($val eq 'ungraded_attempted') {$code = '#';} else {$code = ' ';} $partData{$part.':code'}=$code; } } foreach my $part (split(/\:/,$data->{$sequence.':'.$problemID. ':parts'})) { $output->{$name.':'.$problemID.':'.$part.':wrong'} = $partData{$part.':tries'}; $allkeys{$name.':'.$problemID.':'.$part.':wrong'}++; if($partData{$part.':code'} eq '*') { $output->{$name.':'.$problemID.':'.$part.':wrong'}--; $problemsCorrect++; } elsif($partData{$part.':code'} eq '+') { $output->{$name.':'.$problemID.':'.$part.':wrong'}--; $problemsCorrect++; } $output->{$name.':'.$problemID.':'.$part.':tries'} = $partData{$part.':tries'}; $output->{$name.':'.$problemID.':'.$part.':code'} = $partData{$part.':code'}; $output->{$name.':'.$problemID.':'.$part.':awarded'} = $partData{$part.':awarded'}; $allkeys{$name.':'.$problemID.':'.$part.':tries'}++; $allkeys{$name.':'.$problemID.':'.$part.':code'}++; $allkeys{$name.':'.$problemID.':'.$part.':awarded'}++; $totalAwarded += $partData{$part.':awarded'}; $output->{$name.':'.$problemID.':'.$part.':timestamp'} = $partData{$part.':timestamp'}; $allkeys{$name.':'.$problemID.':'.$part.':timestamp'}++; foreach my $response (split(':', $data->{$sequence.':'. $problemID.':'. $part.':responseIDs'})) { $output->{$name.':'.$problemID.':'.$part.':'.$response. ':submission'}=join(':::',@submissions); $allkeys{$name.':'.$problemID.':'.$part.':'.$response. ':submission'}++; } if($partData{$part.':code'} ne 'x') { $totalProblems++; } } } $output->{$name.':'.$sequence.':problemsCorrect'} = $problemsCorrect; $allkeys{$name.':'.$sequence.':problemsCorrect'}++; $problemsSolved += $problemsCorrect; $problemsCorrect=0; } $output->{$name.':problemsSolved'} = $problemsSolved; $output->{$name.':totalProblems'} = $totalProblems; $output->{$name.':totalAwarded'} = $totalAwarded; $allkeys{$name.':problemsSolved'}++; $allkeys{$name.':totalProblems'}++; $allkeys{$name.':totalAwarded'}++; $output->{$name.':keys'} = join(':::', keys(%allkeys)); return; } sub LoadDiscussion { my ($courseID)=@_; my %Discuss=(); my %contrib=&Apache::lonnet::dump( $courseID, $ENV{'course.'.$courseID.'.domain'}, $ENV{'course.'.$courseID.'.num'}); #my %contrib=&DownloadCourseInformation($name, $courseID, 0); foreach my $temp(keys %contrib) { if ($temp=~/^version/) { my $ver=$contrib{$temp}; my ($dummy,$prb)=split(':',$temp); for (my $idx=1; $idx<=$ver; $idx++ ) { my $name=$contrib{"$idx:$prb:sendername"}; $Discuss{"$name:$prb"}=$idx; } } } return \%Discuss; } # ----- END PROCESSING FUNCTIONS --------------------------------------- =pod =head1 HELPER FUNCTIONS These are just a couple of functions do various odd and end jobs. There was also a couple of bulk functions added. These are &DownloadStudentCourseData(), &DownloadStudentCourseDataSeparate(), and &CheckForResidualDownload(). These functions now act as the interface for downloading student course data. The statistical modules should no longer make the calls to dump and download and process etc. They make calls to these bulk functions to get their data. =cut # ----- HELPER FUNCTIONS ----------------------------------------------- sub CheckDateStampError { my ($courseData, $cache, $name)=@_; if($courseData->{$name.':UpToDate'} eq 'true') { $cache->{$name.':lastDownloadTime'} = $courseData->{$name.':lastDownloadTime'}; if($courseData->{$name.':lastDownloadTime'} eq 'Not downloaded') { $cache->{$name.':updateTime'} = ' Not updated'; } else { $cache->{$name.':updateTime'}= localtime($courseData->{$name.':lastDownloadTime'}); } return 0; } $cache->{$name.':lastDownloadTime'}=$courseData->{$name.':lastDownloadTime'}; if($courseData->{$name.':lastDownloadTime'} eq 'Not downloaded') { $cache->{$name.':updateTime'} = ' Not updated'; } else { $cache->{$name.':updateTime'}= localtime($courseData->{$name.':lastDownloadTime'}); } if(defined($courseData->{$name.':error'})) { $cache->{$name.':error'}=$courseData->{$name.':error'}; return 0; } return 1; } =pod =item &ProcessFullName() Takes lastname, generation, firstname, and middlename (or some partial set of this data) and returns the full name version as a string. Format is Lastname generation, firstname middlename or a subset of this. =cut sub ProcessFullName { my ($lastname, $generation, $firstname, $middlename)=@_; my $Str = ''; # Strip whitespace preceeding & following name components. $lastname =~ s/(\s+$|^\s+)//g; $generation =~ s/(\s+$|^\s+)//g; $firstname =~ s/(\s+$|^\s+)//g; $middlename =~ s/(\s+$|^\s+)//g; if($lastname ne '') { $Str .= $lastname; $Str .= ' '.$generation if ($generation ne ''); $Str .= ','; $Str .= ' '.$firstname if ($firstname ne ''); $Str .= ' '.$middlename if ($middlename ne ''); } else { $Str .= $firstname if ($firstname ne ''); $Str .= ' '.$middlename if ($middlename ne ''); $Str .= ' '.$generation if ($generation ne ''); } return $Str; } =pod =item &TestCacheData() Determine if the cache database can be accessed with a tie. It waits up to ten seconds before returning failure. This function exists to help with the problems with stopping the data download. When an abort occurs and the user quickly presses a form button and httpd child is created. This child needs to wait for the other to finish (hopefully within ten seconds). =over 4 Input: $ChartDB $ChartDB: The name of the cache database to be opened Output: -1, 0, 1 -1: Could not tie database 0: Use cached data 1: New cache database created, use that. =back =cut sub TestCacheData { my ($ChartDB,$isRecalculate,$totalDelay)=@_; my $isCached=-1; my %testData; my $tieTries=0; if(!defined($totalDelay)) { $totalDelay = 10; } if ((-e "$ChartDB") && (!$isRecalculate)) { $isCached = 1; } else { $isCached = 0; } while($tieTries < $totalDelay) { my $result=0; if($isCached) { $result=tie(%testData,'GDBM_File',$ChartDB,&GDBM_READER(),0640); } else { $result=tie(%testData,'GDBM_File',$ChartDB,&GDBM_NEWDB(),0640); } if($result) { last; } $tieTries++; sleep 1; } if($tieTries >= $totalDelay) { return -1; } untie(%testData); return $isCached; } sub DownloadStudentCourseData { my ($students,$checkDate,$cacheDB,$extract,$status,$courseID,$r,$c)=@_; my $title = 'LON-CAPA Statistics'; my $heading = 'Download and Process Course Data'; my $studentCount = scalar(@$students); my $WhatIWant; $WhatIWant = '(^version:|'; $WhatIWant .= '^\d+:.+?:(resource\.\d+\.'; $WhatIWant .= '(solved|tries|previous|awarded|(\d+\.submission))\s*$';#' $WhatIWant .= '|timestamp)'; $WhatIWant .= ')'; # $WhatIWant = '.'; my %prog_state; if($status eq 'true') { %prog_state=&Apache::lonhtmlcommon::Create_PrgWin($r, $title, $heading,($#$students)+1); } foreach (@$students) { my %cache; if($c->aborted()) { return 'Aborted'; } if($status eq 'true') { &Apache::lonhtmlcommon::Increment_PrgWin($r,\%prog_state, 'last student '.$_); } my $downloadTime='Not downloaded'; my $needUpdate = 'false'; if($checkDate eq 'true' && tie(%cache,'GDBM_File',$cacheDB,&GDBM_READER(),0640)) { $downloadTime = $cache{$_.':lastDownloadTime'}; $needUpdate = $cache{'ResourceUpdated'}; untie(%cache); } if($c->aborted()) { return 'Aborted'; } if($needUpdate eq 'true') { $downloadTime = 'Not downloaded'; } my $courseData = &DownloadCourseInformation($_, $courseID, $downloadTime, $WhatIWant); if(tie(%cache,'GDBM_File',$cacheDB,&GDBM_WRCREAT(),0640)) { foreach my $key (keys(%$courseData)) { if($key =~ /^(con_lost|error|no_such_host)/i) { $courseData->{$_.':error'} = 'No course data for '.$_; last; } } if($extract eq 'true') { &ExtractStudentData($courseData, \%cache, \%cache, $_); } else { &ProcessStudentData(\%cache, $courseData, $_); } untie(%cache); } else { next; } } if($status eq 'true') { &Apache::lonhtmlcommon::Close_PrgWin($r,\%prog_state); } return 'OK'; } sub DownloadStudentCourseDataSeparate { my ($students,$checkDate,$cacheDB,$extract,$status,$courseID,$r,$c)=@_; my $residualFile = $Apache::lonnet::tmpdir.$courseID.'DownloadFile.db'; my $title = 'LON-CAPA Statistics'; my $heading = 'Download Course Data'; my $WhatIWant; $WhatIWant = '(^version:|'; $WhatIWant .= '^\d+:.+?:(resource\.\d+\.'; $WhatIWant .= '(solved|tries|previous|awarded|(\d+\.submission))\s*$';#' $WhatIWant .= '|timestamp)'; $WhatIWant .= ')'; &CheckForResidualDownload($cacheDB, 'true', 'true', $courseID, $r, $c); my $studentCount = scalar(@$students); my %prog_state; if($status eq 'true') { %prog_state=&Apache::lonhtmlcommon::Create_PrgWin($r, $title, $heading,($#$students)+1); } my $displayString=''; foreach (@$students) { if($c->aborted()) { return 'Aborted'; } if($status eq 'true') { &Apache::lonhtmlcommon::Increment_PrgWin($r,\%prog_state, 'last student '.$_); } my %cache; my $downloadTime='Not downloaded'; my $needUpdate = 'false'; if($checkDate eq 'true' && tie(%cache,'GDBM_File',$cacheDB,&GDBM_READER(),0640)) { $downloadTime = $cache{$_.':lastDownloadTime'}; $needUpdate = $cache{'ResourceUpdated'}; untie(%cache); } if($c->aborted()) { return 'Aborted'; } if($needUpdate eq 'true') { $downloadTime = 'Not downloaded'; } my $error = 0; my $courseData = &DownloadCourseInformation($_, $courseID, $downloadTime, $WhatIWant); my %downloadData; unless(tie(%downloadData,'GDBM_File',$residualFile, &GDBM_WRCREAT(),0640)) { return 'Failed to tie temporary download hash.'; } foreach my $key (keys(%$courseData)) { $downloadData{$key} = $courseData->{$key}; if($key =~ /^(con_lost|error|no_such_host)/i) { $error = 1; last; } } if($error) { foreach my $deleteKey (keys(%$courseData)) { delete $downloadData{$deleteKey}; } $downloadData{$_.':error'} = 'No course data for '.$_; } untie(%downloadData); } if($status eq 'true') { &Apache::lonhtmlcommon::Close_PrgWin($r, \%prog_state); } return &CheckForResidualDownload($cacheDB, 'true', 'true', $courseID, $r, $c); } sub CheckForResidualDownload { my ($cacheDB,$extract,$status,$courseID,$r,$c)=@_; my $residualFile = $Apache::lonnet::tmpdir.$courseID.'DownloadFile.db'; if(!-e $residualFile) { return 'OK'; } my %downloadData; my %cache; unless(tie(%downloadData,'GDBM_File',$residualFile,&GDBM_READER(),0640)) { return 'Can not tie database for check for residual download: tempDB'; } unless(tie(%cache,'GDBM_File',$cacheDB,&GDBM_WRCREAT(),0640)) { untie(%downloadData); return 'Can not tie database for check for residual download: cacheDB'; } my @students=(); my %checkStudent; my $key; while(($key, undef) = each %downloadData) { my @temp = split(':', $key); my $student = $temp[0].':'.$temp[1]; if(!defined($checkStudent{$student})) { $checkStudent{$student}++; push(@students, $student); } } my $heading = 'Process Course Data'; my $title = 'LON-CAPA Statistics'; my $studentCount = scalar(@students); my %prog_state; if($status eq 'true') { %prog_state=&Apache::lonhtmlcommon::Create_PrgWin($r, $title, $heading,$#students+1); } my $count=1; foreach my $name (@students) { last if($c->aborted()); if($status eq 'true') { &Apache::lonhtmlcommon::Increment_PrgWin($r,\%prog_state, 'last student '.$name); } if($extract eq 'true') { &ExtractStudentData(\%downloadData, \%cache, \%cache, $name); } else { &ProcessStudentData(\%cache, \%downloadData, $name); } $count++; } if($status eq 'true') { &Apache::lonhtmlcommon::Close_PrgWin($r, \%prog_state); } untie(%cache); untie(%downloadData); if(!$c->aborted()) { my @files = ($residualFile); unlink(@files); } return 'OK'; } ################################################ ################################################ =pod =item &make_into_hash($values); Returns a reference to a hash as described by $values. $values is assumed to be the result of join(':',map {&Apache::lonnet::escape($_)} %orighash); This is a helper function for get_current_state. =cut ################################################ ################################################ sub make_into_hash { my $values = shift; my %tmp = map { &Apache::lonnet::unescape($_); } split(':',$values); return \%tmp; } ################################################ ################################################ =pod =head1 LOCAL DATA CACHING SUBROUTINES The local caching is done using MySQL. There is no fall-back implementation if MySQL is not running. The programmers interface is to call &get_current_state() or some other primary interface subroutine (described below). The internals of this storage system are documented here. There are six tables used to store student performance data (the results of a dumpcurrent). Each of these tables is created in MySQL with a name of $courseid_*****, where ***** is 'symb', 'part', or whatever is appropriate for the table. The tables and their purposes are described below. Some notes before we get started. Each table must have a PRIMARY KEY, which is a column or set of columns which will serve to uniquely identify a row of data. NULL is not allowed! INDEXes work best on integer data. JOIN is used to combine data from many tables into one output. lonmysql.pm is used for some of the interface, specifically the table creation calls. The inserts are done in bulk by directly calling the database handler. The SELECT ... JOIN statement used to retrieve the data does not have an interface in lonmysql.pm and I shudder at the thought of writing one. =head3 Table Descriptions =over 4 =item $symb_table The symb_table has two columns. The first is a 'symb_id' and the second is the text name for the 'symb' (limited to 64k). The 'symb_id' is generated automatically by MySQL so inserts should be done on this table with an empty first element. This table has its PRIMARY KEY on the 'symb_id'. =item $part_table The part_table has two columns. The first is a 'part_id' and the second is the text name for the 'part' (limited to 100 characters). The 'part_id' is generated automatically by MySQL so inserts should be done on this table with an empty first element. This table has its PRIMARY KEY on the 'part' (100 characters) and a KEY on 'part_id'. =item $student_table The student_table has two columns. The first is a 'student_id' and the second is the text description of the 'student' (typically username:domain) (less than 100 characters). The 'student_id' is automatically generated by MySQL. The use of the name 'student_id' is loaded, I know, but this ID is used ONLY internally to the MySQL database and is not the same as the students ID (stored in the students environment). This table has its PRIMARY KEY on the 'student' (100 characters). =item $updatetime_table The updatetime_table has two columns. The first is 'student' (100 characters, typically username:domain). The second is 'updatetime', which is an unsigned integer, NOT a MySQL date. This table has its PRIMARY KEY on 'student' (100 characters). =item $performance_table The performance_table has 9 columns. The first three are 'symb_id', 'student_id', and 'part_id'. These comprise the PRIMARY KEY for this table and are directly related to the $symb_table, $student_table, and $part_table described above. MySQL does better indexing on numeric items than text, so we use these three "index tables". The remaining columns are 'solved', 'tries', 'awarded', 'award', 'awarddetail', and 'timestamp'. These are either the MySQL type TINYTEXT or various integers ('tries' and 'timestamp'). This table has KEYs of 'student_id' and 'symb_id'. For use of this table, see the functions described below. =item $parameters_table The parameters_table holds the data that does not fit neatly into the performance_table. The parameters table has four columns: 'symb_id', 'student_id', 'parameter', and 'value'. 'symb_id', 'student_id', and 'parameter' comprise the PRIMARY KEY for this table. 'parameter' is limited to 255 characters. 'value' is limited to 64k characters. =back =head3 Important Subroutines Here is a brief overview of the subroutines which are likely to be of interest: =over 4 =item &get_current_state(): programmers interface. =item &init_dbs(): table creation =item &update_student_data(): data storage calls =item &get_student_data_from_performance_cache(): data retrieval =back =head3 Main Documentation =over 4 =cut ################################################ ################################################ ################################################ ################################################ { my $current_course =''; my $symb_table; my $part_table; my $student_table; my $updatetime_table; my $performance_table; my $parameters_table; ################################################ ################################################ =pod =item &setup_table_names() input: course id output: none Sets the package variables for the MySQL table names: =over 4 =item $symb_table =item $part_table =item $student_table =item $updatetime_table =item $performance_table =item $parameters_table =back =cut ################################################ ################################################ sub setup_table_names { my $courseid = shift; if (! defined($courseid)) { $courseid = $ENV{'request.course.id'}; } # # Set up database names my $base_id = $courseid; $symb_table = $base_id.'_'.'symb'; $part_table = $base_id.'_'.'part'; $student_table = $base_id.'_'.'student'; $updatetime_table = $base_id.'_'.'updatetime'; $performance_table = $base_id.'_'.'performance'; $parameters_table = $base_id.'_'.'parameters'; return; } ################################################ ################################################ =pod =item &init_dbs() Input: course id Output: 0 on success, positive integer on error This routine issues the calls to lonmysql to create the tables used to store student data. =cut ################################################ ################################################ sub init_dbs { my $courseid = shift; &setup_table_names($courseid); # # Note - changes to this table must be reflected in the code that # stores the data (calls &Apache::lonmysql::store_row with this table # id my $symb_table_def = { id => $symb_table, permanent => 'no', columns => [{ name => 'symb_id', type => 'MEDIUMINT UNSIGNED', restrictions => 'NOT NULL', auto_inc => 'yes', }, { name => 'symb', type => 'MEDIUMTEXT', restrictions => 'NOT NULL'}, ], 'PRIMARY KEY' => ['symb_id'], }; # my $part_table_def = { id => $part_table, permanent => 'no', columns => [{ name => 'part_id', type => 'MEDIUMINT UNSIGNED', restrictions => 'NOT NULL', auto_inc => 'yes', }, { name => 'part', type => 'VARCHAR(100)', restrictions => 'NOT NULL'}, ], 'PRIMARY KEY' => ['part (100)'], 'KEY' => [{ columns => ['part_id']},], }; # my $student_table_def = { id => $student_table, permanent => 'no', columns => [{ name => 'student_id', type => 'MEDIUMINT UNSIGNED', restrictions => 'NOT NULL', auto_inc => 'yes', }, { name => 'student', type => 'VARCHAR(100)', restrictions => 'NOT NULL'}, ], 'PRIMARY KEY' => ['student (100)'], 'KEY' => [{ columns => ['student_id']},], }; # my $updatetime_table_def = { id => $updatetime_table, permanent => 'no', columns => [{ name => 'student', type => 'VARCHAR(100)', restrictions => 'NOT NULL UNIQUE',}, { name => 'updatetime', type => 'INT UNSIGNED', restrictions => 'NOT NULL' }, ], 'PRIMARY KEY' => ['student (100)'], }; # my $performance_table_def = { id => $performance_table, permanent => 'no', columns => [{ name => 'symb_id', type => 'MEDIUMINT UNSIGNED', restrictions => 'NOT NULL' }, { name => 'student_id', type => 'MEDIUMINT UNSIGNED', restrictions => 'NOT NULL' }, { name => 'part_id', type => 'MEDIUMINT UNSIGNED', restrictions => 'NOT NULL' }, { name => 'solved', type => 'TINYTEXT' }, { name => 'tries', type => 'SMALLINT UNSIGNED' }, { name => 'awarded', type => 'TINYTEXT' }, { name => 'award', type => 'TINYTEXT' }, { name => 'awarddetail', type => 'TINYTEXT' }, { name => 'timestamp', type => 'INT UNSIGNED'}, ], 'PRIMARY KEY' => ['symb_id','student_id','part_id'], 'KEY' => [{ columns=>['student_id'] }, { columns=>['symb_id'] },], }; # my $parameters_table_def = { id => $parameters_table, permanent => 'no', columns => [{ name => 'symb_id', type => 'MEDIUMINT UNSIGNED', restrictions => 'NOT NULL' }, { name => 'student_id', type => 'MEDIUMINT UNSIGNED', restrictions => 'NOT NULL' }, { name => 'parameter', type => 'TINYTEXT', restrictions => 'NOT NULL' }, { name => 'value', type => 'MEDIUMTEXT' }, ], 'PRIMARY KEY' => ['symb_id','student_id','parameter (255)'], }; # # Create the tables my $tableid; $tableid = &Apache::lonmysql::create_table($symb_table_def); if (! defined($tableid)) { &Apache::lonnet::logthis("error creating symb_table: ". &Apache::lonmysql::get_error()); return 1; } # $tableid = &Apache::lonmysql::create_table($part_table_def); if (! defined($tableid)) { &Apache::lonnet::logthis("error creating part_table: ". &Apache::lonmysql::get_error()); return 2; } # $tableid = &Apache::lonmysql::create_table($student_table_def); if (! defined($tableid)) { &Apache::lonnet::logthis("error creating student_table: ". &Apache::lonmysql::get_error()); return 3; } # $tableid = &Apache::lonmysql::create_table($updatetime_table_def); if (! defined($tableid)) { &Apache::lonnet::logthis("error creating updatetime_table: ". &Apache::lonmysql::get_error()); return 4; } # $tableid = &Apache::lonmysql::create_table($performance_table_def); if (! defined($tableid)) { &Apache::lonnet::logthis("error creating preformance_table: ". &Apache::lonmysql::get_error()); return 5; } # $tableid = &Apache::lonmysql::create_table($parameters_table_def); if (! defined($tableid)) { &Apache::lonnet::logthis("error creating parameters_table: ". &Apache::lonmysql::get_error()); return 6; } return 0; } ################################################ ################################################ =pod =item &get_part_id() Get the MySQL id of a problem part string. Input: $part Output: undef on error, integer $part_id on success. =item &get_part() Get the string describing a part from the MySQL id of the problem part. Input: $part_id Output: undef on error, $part string on success. =cut ################################################ ################################################ my %ids_by_part; my %parts_by_id; sub get_part_id { my ($part) = @_; if (! exists($ids_by_part{$part})) { &Apache::lonmysql::store_row($part_table,[undef,$part]); undef(%ids_by_part); my @Result = &Apache::lonmysql::get_rows($part_table); foreach (@Result) { $ids_by_part{$_->[1]}=$_->[0]; } } return $ids_by_part{$part} if (exists($ids_by_part{$part})); return undef; # error } sub get_part { my ($part_id) = @_; if (! exists($parts_by_id{$part_id}) || ! defined($parts_by_id{$part_id}) || $parts_by_id{$part_id} eq '') { my @Result = &Apache::lonmysql::get_rows($part_table); foreach (@Result) { $parts_by_id{$_->[0]}=$_->[1]; } } return $parts_by_id{$part_id} if(exists($parts_by_id{$part_id})); return undef; # error } ################################################ ################################################ =pod =item &get_symb_id() Get the MySQL id of a symb. Input: $symb Output: undef on error, integer $symb_id on success. =item &get_symb() Get the symb associated with a MySQL symb_id. Input: $symb_id Output: undef on error, $symb on success. =cut ################################################ ################################################ my %ids_by_symb; my %symbs_by_id; sub get_symb_id { my ($symb) = @_; if (! exists($ids_by_symb{$symb})) { &Apache::lonmysql::store_row($symb_table,[undef,$symb]); undef(%ids_by_symb); my @Result = &Apache::lonmysql::get_rows($symb_table); foreach (@Result) { $ids_by_symb{$_->[1]}=$_->[0]; } } return $ids_by_symb{$symb} if(exists( $ids_by_symb{$symb})); return undef; # error } sub get_symb { my ($symb_id) = @_; if (! exists($symbs_by_id{$symb_id}) || ! defined($symbs_by_id{$symb_id}) || $symbs_by_id{$symb_id} eq '') { my @Result = &Apache::lonmysql::get_rows($symb_table); foreach (@Result) { $symbs_by_id{$_->[0]}=$_->[1]; } } return $symbs_by_id{$symb_id} if(exists( $symbs_by_id{$symb_id})); return undef; # error } ################################################ ################################################ =pod =item &get_student_id() Get the MySQL id of a student. Input: $sname, $dom Output: undef on error, integer $student_id on success. =item &get_student() Get student username:domain associated with the MySQL student_id. Input: $student_id Output: undef on error, string $student (username:domain) on success. =cut ################################################ ################################################ my %ids_by_student; my %students_by_id; sub get_student_id { my ($sname,$sdom) = @_; my $student = $sname.':'.$sdom; if (! exists($ids_by_student{$student})) { &Apache::lonmysql::store_row($student_table,[undef,$student]); undef(%ids_by_student); my @Result = &Apache::lonmysql::get_rows($student_table); foreach (@Result) { $ids_by_student{$_->[1]}=$_->[0]; } } return $ids_by_student{$student} if(exists( $ids_by_student{$student})); return undef; # error } sub get_student { my ($student_id) = @_; if (! exists($students_by_id{$student_id}) || ! defined($students_by_id{$student_id}) || $students_by_id{$student_id} eq '') { my @Result = &Apache::lonmysql::get_rows($student_table); foreach (@Result) { $students_by_id{$_->[0]}=$_->[1]; } } return $students_by_id{$student_id} if(exists($students_by_id{$student_id})); return undef; # error } ################################################ ################################################ =pod =item &update_student_data() Input: $sname, $sdom, $courseid Output: $returnstatus, \%student_data $returnstatus is a string describing any errors that occured. 'okay' is the default. \%student_data is the data returned by a call to lonnet::currentdump. This subroutine loads a students data using lonnet::currentdump and inserts it into the MySQL database. The inserts are done on two tables, $performance_table and $parameters_table. $parameters_table holds the data that is not included in $performance_table. See the description of $performance_table elsewhere in this file. The INSERT calls are made directly by this subroutine, not through lonmysql because we do a 'bulk' insert which takes advantage of MySQLs non-SQL compliant INSERT command to insert multiple rows at a time. If anything has gone wrong during this process, $returnstatus is updated with a description of the error and \%student_data is returned. Notice we do not insert the data and immediately query it. This means it is possible for there to be data returned this first time that is not available the second time. CYA. =cut ################################################ ################################################ sub update_student_data { my ($sname,$sdom,$courseid) = @_; # my $student_id = &get_student_id($sname,$sdom); my $student = $sname.':'.$sdom; # my $returnstatus = 'okay'; # # Set up database names &setup_table_names($courseid); # # Download students data my $time_of_retrieval = time; my @tmp = &Apache::lonnet::currentdump($courseid,$sdom,$sname); if ((scalar(@tmp) > 0) && ($tmp[0] =~ /^error:/)) { &Apache::lonnet::logthis('error getting data for '. $sname.':'.$sdom.' in course '.$courseid. ':'.$tmp[0]); $returnstatus = 'error getting data'; return $returnstatus; } if (scalar(@tmp) < 1) { return ('no data',undef); } my %student_data = @tmp; # # Remove all of the students data from the table &Apache::lonmysql::remove_from_table($performance_table,'student_id', $student_id); # # Store away the data # my $starttime = Time::HiRes::time; my $elapsed = 0; my $rows_stored; my $store_parameters_command = 'INSERT INTO '.$parameters_table. ' VALUES '; my $store_performance_command = 'INSERT INTO '.$performance_table. ' VALUES '; my $dbh = &Apache::lonmysql::get_dbh(); return 'error' if (! defined($dbh)); while (my ($current_symb,$param_hash) = each(%student_data)) { # # make sure the symb is set up properly my $symb_id = &get_symb_id($current_symb); # # Load data into the tables while (my ($parameter,$value) = each (%$param_hash)) { my $newstring; if ($parameter !~ /(timestamp|resource\.(.*)\.(solved|tries|awarded|award|awarddetail|previous))/) { $newstring = "('".join("','", $symb_id,$student_id, $parameter,$value)."'),"; if ($newstring !~ /''/) { $store_parameters_command .= $newstring; $rows_stored++; } } next if ($parameter !~ /^resource\.(.*)\.solved$/); # my $part = $1; my $part_id = &get_part_id($part); next if (!defined($part_id)); my $solved = $value; my $tries = $param_hash->{'resource.'.$part.'.tries'}; my $awarded = $param_hash->{'resource.'.$part.'.awarded'}; my $award = $param_hash->{'resource.'.$part.'.award'}; my $awarddetail = $param_hash->{'resource.'.$part.'.awarddetail'}; my $timestamp = $param_hash->{'timestamp'}; $solved = '' if (! defined($awarded)); $tries = '' if (! defined($tries)); $awarded = '' if (! defined($awarded)); $award = '' if (! defined($award)); $awarddetail = '' if (! defined($awarddetail)); $newstring = "('".join("','",$symb_id,$student_id,$part_id, $solved,$tries,$awarded,$award, $awarddetail,$timestamp)."'),"; $store_performance_command .= $newstring; $rows_stored++; } } chop $store_parameters_command; chop $store_performance_command; my $start = Time::HiRes::time; $dbh->do($store_parameters_command); if ($dbh->err()) { &Apache::lonnet::logthis(' bigass insert error:'.$dbh->errstr()); &Apache::lonnet::logthis('command = '.$store_performance_command); $returnstatus = 'error: unable to insert parameters into database'; return $returnstatus,\%student_data; } $dbh->do($store_performance_command); if ($dbh->err()) { &Apache::lonnet::logthis(' bigass insert error:'.$dbh->errstr()); &Apache::lonnet::logthis('command = '.$store_parameters_command); $returnstatus = 'error: unable to insert performance into database'; return $returnstatus,\%student_data; } $elapsed += Time::HiRes::time - $start; # # Set the students update time &Apache::lonmysql::replace_row($updatetime_table, [$student,$time_of_retrieval]); &Apache::lonnet::logthis('store took: '.(Time::HiRes::time - $starttime).' for '.$rows_stored); &Apache::lonnet::logthis('mysql store took: '.$elapsed.' for '.$rows_stored); return ($returnstatus,\%student_data); } ################################################ ################################################ =pod =item &ensure_current_data() Input: $sname, $sdom, $courseid Output: $status, $data This routine ensures the data for a given student is up to date. It calls &init_dbs() if the tables do not exist. The $updatetime_table is queried to determine the time of the last update. If the students data is out of date, &update_student_data() is called. The return values from the call to &update_student_data() are returned. =cut ################################################ ################################################ sub ensure_current_data { my ($sname,$sdom,$courseid) = @_; my $status = 'okay'; # return value # &setup_table_names($courseid); # # if the tables do not exist, make them my @CurrentTable = &Apache::lonmysql::tables_in_db(); my ($found_symb,$found_student,$found_part,$found_update, $found_performance,$found_parameters); foreach (@CurrentTable) { $found_symb = 1 if ($_ eq $symb_table); $found_student = 1 if ($_ eq $student_table); $found_part = 1 if ($_ eq $part_table); $found_update = 1 if ($_ eq $updatetime_table); $found_performance = 1 if ($_ eq $performance_table); $found_parameters = 1 if ($_ eq $parameters_table); } if (!$found_symb || !$found_update || !$found_student || !$found_part || !$found_performance || !$found_parameters) { if (&init_dbs($courseid)) { return 'error'; } } # # Get the update time for the user my $updatetime = 0; my $modifiedtime = 1; # my $student = $sname.':'.$sdom; my @Result = &Apache::lonmysql::get_rows($updatetime_table, "student ='$student'"); my $data = undef; if (@Result) { $updatetime = $Result[0]->[1]; } if ($modifiedtime > $updatetime) { ($status,$data) = &update_student_data($sname,$sdom,$courseid); } return ($status,$data); } ################################################ ################################################ =pod =item &get_student_data_from_performance_cache() Input: $sname, $sdom, $symb, $courseid Output: hash reference containing the data for the given student. If $symb is undef, all the students data is returned. This routine is the heart of the local caching system. See the description of $performance_table, $symb_table, $student_table, and $part_table. The main task is building the MySQL request. The tables appear in the request in the order in which they should be parsed by MySQL. When searching on a student the $student_table is used to locate the 'student_id'. All rows in $performance_table which have a matching 'student_id' are returned, with data from $part_table and $symb_table which match the entries in $performance_table, 'part_id' and 'symb_id'. When searching on a symb, the $symb_table is processed first, with matching rows grabbed from $performance_table and filled in from $part_table and $student_table in that order. Running 'EXPLAIN ' on the 'SELECT' statements generated can be quite interesting, especially if you play with the order the tables are listed. =cut ################################################ ################################################ sub get_student_data_from_performance_cache { my ($sname,$sdom,$symb,$courseid)=@_; my $student = $sname.':'.$sdom if (defined($sname) && defined($sdom)); &setup_table_names(); # # Return hash my $studentdata; # my $dbh = &Apache::lonmysql::get_dbh(); my $request = "SELECT ". "d.symb,c.part,a.solved,a.tries,a.awarded,a.award,a.awarddetail,". "a.timestamp "; if (defined($student)) { $request .= "FROM $student_table AS b ". "LEFT JOIN $performance_table AS a ON b.student_id=a.student_id ". "LEFT JOIN $part_table AS c ON c.part_id = a.part_id ". "LEFT JOIN $symb_table AS d ON d.symb_id = a.symb_id ". "WHERE student='$student'"; if (defined($symb) && $symb ne '') { $request .= " AND d.symb='".$dbh->quote($symb)."'"; } } elsif (defined($symb) && $symb ne '') { $request .= "FROM $symb_table as d ". "LEFT JOIN $performance_table AS a ON d.symb_id=a.symb_id ". "LEFT JOIN $part_table AS c ON c.part_id = a.part_id ". "LEFT JOIN $student_table AS b ON b.student_id = a.student_id ". "WHERE symb='".$dbh->quote($symb)."'"; } my $starttime = Time::HiRes::time; my $rows_retrieved = 0; my $sth = $dbh->prepare($request); $sth->execute(); if ($sth->err()) { &Apache::lonnet::logthis("Unable to execute MySQL request:"); &Apache::lonnet::logthis("\n".$request."\n"); &Apache::lonnet::logthis("error is:".$sth->errstr()); return undef; } foreach my $row (@{$sth->fetchall_arrayref}) { $rows_retrieved++; my ($symb,$part,$solved,$tries,$awarded,$award,$awarddetail,$time) = (@$row); my $base = 'resource.'.$part; $studentdata->{$symb}->{$base.'.solved'} = $solved; $studentdata->{$symb}->{$base.'.tries'} = $tries; $studentdata->{$symb}->{$base.'.awarded'} = $awarded; $studentdata->{$symb}->{$base.'.award'} = $award; $studentdata->{$symb}->{$base.'.awarddetail'} = $awarddetail; $studentdata->{$symb}->{'timestamp'} = $time if (defined($time) && $time ne ''); } &Apache::lonnet::logthis('retrieve took: '.(Time::HiRes::time - $starttime).' for '.$rows_retrieved); return $studentdata; } ################################################ ################################################ =pod =item &get_current_state() Input: $sname,$sdom,$symb,$courseid Output: Described below Retrieve the current status of a students performance. $sname and $sdom are the only required parameters. If $symb is undef the results of an &Apache::lonnet::currentdump() will be returned. If $courseid is undef it will be retrieved from the environment. The return structure is based on &Apache::lonnet::currentdump. If $symb is unspecified, all the students data is returned in a hash of the form: ( symb1 => { param1 => value1, param2 => value2 ... }, symb2 => { param1 => value1, param2 => value2 ... }, ) If $symb is specified, a hash of ( param1 => value1, param2 => value2, ) is returned. If no data is found for $symb, or if the student has no performance data, an empty list is returned. =cut ################################################ ################################################ sub get_current_state { my ($sname,$sdom,$symb,$courseid,$forcedownload)=@_; if ($current_course ne $courseid) { # Clear out variables undef(%ids_by_part); undef(%parts_by_id); undef(%ids_by_symb); undef(%symbs_by_id); undef(%ids_by_student); undef(%students_by_id); $current_course = $courseid; } return () if (! defined($sname) || ! defined($sdom)); # $courseid = $ENV{'request.course.id'} if (! defined($courseid)); # my ($status,$data) = &ensure_current_data($sname,$sdom,$courseid); # if (defined($data)) { return %$data; } elsif ($status eq 'no data') { return (); } else { if ($status ne 'okay' && $status ne '') { &Apache::lonnet::logthis('status = '.$status); return (); } my $returnhash = &get_student_data_from_performance_cache($sname,$sdom, $symb,$courseid); return %$returnhash if (defined($returnhash)); } return (); } ################################################ ################################################ =pod =back =item End of Local Data Caching Subroutines =cut ################################################ ################################################ } ################################################ ################################################ =pod =head3 Classlist Subroutines =item &get_classlist(); Retrieve the classist of a given class or of the current class. Student information is returned from the classlist.db file and, if needed, from the students environment. Optional arguments are $cid, $cdom, and $cnum (course id, course domain, and course number, respectively). Any omitted arguments will be taken from the current environment ($ENV{'request.course.id'}, $ENV{'course.'.$cid.'.domain'}, and $ENV{'course.'.$cid.'.num'}). Returns a reference to a hash which contains: keys '$sname:$sdom' values [$sdom,$sname,$end,$start,$id,$section,$fullname,$status] The constant values CL_SDOM, CL_SNAME, CL_END, etc. can be used as indices into the returned list to future-proof clients against changes in the list order. =cut ################################################ ################################################ sub CL_SDOM { return 0; } sub CL_SNAME { return 1; } sub CL_END { return 2; } sub CL_START { return 3; } sub CL_ID { return 4; } sub CL_SECTION { return 5; } sub CL_FULLNAME { return 6; } sub CL_STATUS { return 7; } sub get_classlist { my ($cid,$cdom,$cnum) = @_; $cid = $cid || $ENV{'request.course.id'}; $cdom = $cdom || $ENV{'course.'.$cid.'.domain'}; $cnum = $cnum || $ENV{'course.'.$cid.'.num'}; my $now = time; # my %classlist=&Apache::lonnet::dump('classlist',$cdom,$cnum); while (my ($student,$info) = each(%classlist)) { return undef if ($student =~ /^(con_lost|error|no_such_host)/i); my ($sname,$sdom) = split(/:/,$student); my @Values = split(/:/,$info); my ($end,$start,$id,$section,$fullname); if (@Values > 2) { ($end,$start,$id,$section,$fullname) = @Values; } else { # We have to get the data ourselves ($end,$start) = @Values; $section = &Apache::lonnet::getsection($sdom,$sname,$cid); my %info=&Apache::lonnet::get('environment', ['firstname','middlename', 'lastname','generation','id'], $sdom, $sname); my ($tmp) = keys(%info); if ($tmp =~/^(con_lost|error|no_such_host)/i) { $fullname = 'not available'; $id = 'not available'; &Apache::lonnet::logthis('unable to retrieve environment '. 'for '.$sname.':'.$sdom); } else { $fullname = &ProcessFullName(@info{qw/lastname generation firstname middlename/}); $id = $info{'id'}; } # Update the classlist with this students information if ($fullname ne 'not available') { my $enrolldata = join(':',$end,$start,$id,$section,$fullname); my $reply=&Apache::lonnet::cput('classlist', {$student => $enrolldata}, $cdom,$cnum); if ($reply !~ /^(ok|delayed)/) { &Apache::lonnet::logthis('Unable to update classlist for '. 'student '.$sname.':'.$sdom. ' error:'.$reply); } } } my $status='Expired'; if(((!$end) || $now < $end) && ((!$start) || ($now > $start))) { $status='Active'; } $classlist{$student} = [$sdom,$sname,$end,$start,$id,$section,$fullname,$status]; } if (wantarray()) { return (\%classlist,['domain','username','end','start','id', 'section','fullname','status']); } else { return \%classlist; } } # ----- END HELPER FUNCTIONS -------------------------------------------- 1; __END__