[Bio] / Sprout / ERDBLoader.pl Repository:
ViewVC logotype

Annotation of /Sprout/ERDBLoader.pl

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1.8 - (view) (download) (as text)

1 : parrello 1.1 #!/usr/bin/perl -w
2 :    
3 :     #
4 :     # Copyright (c) 2003-2006 University of Chicago and Fellowship
5 :     # for Interpretations of Genomes. All Rights Reserved.
6 :     #
7 :     # This file is part of the SEED Toolkit.
8 :     #
9 :     # The SEED Toolkit is free software. You can redistribute
10 :     # it and/or modify it under the terms of the SEED Toolkit
11 :     # Public License.
12 :     #
13 :     # You should have received a copy of the SEED Toolkit Public License
14 :     # along with this program; if not write to the University of Chicago
15 :     # at info@ci.uchicago.edu or the Fellowship for Interpretation of
16 :     # Genomes at veronika@thefig.info or download a copy from
17 :     # http://www.theseed.org/LICENSE.TXT.
18 :     #
19 :    
20 :     use strict;
21 :     use Tracer;
22 :     use ERDB;
23 :     use ERDBLoadGroup;
24 :     use ERDBGenerate;
25 :     use Stats;
26 : parrello 1.2 use Time::HiRes;
27 : parrello 1.1
28 :    
29 :     =head1 ERDBLoader Script
30 :    
31 :     ERDBLoader [options] <database> <group1> <group2> ...
32 :    
33 :     ERDB Database Load Finisher
34 :    
35 :     =head2 Introduction
36 :    
37 :     This script finishes the database load process begun by [[ERDBGeneratorPl]].
38 :    
39 :     [[ERDBGeneratorPl]] divides the source data into sections, and generates a
40 :     partial load file for each section of each table. To finish the load process, we
41 :     need to combine the partial files into single files and load the resulting
42 :     single files into the database tables.
43 :    
44 :     Like [[ERDBGeneratorPl]], this script acts on load groups-- sets of related
45 :     tables that are loaded at the same time. For each table in a named group that
46 :     does not exist in the database, the script first attempts to find a completed
47 :     data file. If one does not exist, it attempts to create one by collating section
48 : parrello 1.2 files. Once the collated section files for a load group are finished, they are
49 :     loaded into the database.
50 : parrello 1.1
51 :     =head2 Positional Parameters
52 :    
53 :     =over 4
54 :    
55 :     =item database
56 :    
57 :     Name of the ERDB database. This should be the class name for the subclass used
58 :     to access the database.
59 :    
60 :     =back
61 :    
62 :     =head2 Command-Line Options
63 :    
64 :     =over 4
65 :    
66 :     =item trace
67 :    
68 :     Specifies the tracing level. The higher the tracing level, the more messages
69 :     will appear in the trace log. Use E to specify emergency tracing.
70 :    
71 :     =item user
72 :    
73 :     Name suffix to be used for log files. If omitted, the PID is used.
74 :    
75 :     =item sql
76 :    
77 :     If specified, turns on tracing of SQL activity.
78 :    
79 : parrello 1.7 =item clear
80 :    
81 :     If specified, existing load files will be recreated from sections if the sections
82 :     are present.
83 :    
84 : parrello 1.1 =item background
85 :    
86 :     Save the standard and error output to files. The files will be created
87 :     in the FIG temporary directory and will be named C<err>I<User>C<.log> and
88 :     C<out>I<User>C<.log>, respectively, where I<User> is the value of the
89 :     B<user> option above.
90 :    
91 :     =item help
92 :    
93 :     Display this command's parameters and options.
94 :    
95 : parrello 1.3 =item keepSections
96 :    
97 :     If specified, section files (the fragments of data load files created by
98 :     [[ERDBGeneratorPl]], will not be deleted after they are collated.
99 :    
100 : parrello 1.4 =item sanityCheck
101 :    
102 :     If specified, no tables will be loaded. Instead, the first I<N> records from the
103 :     assembled load files will be displayed so that the file contents can be
104 :     visually matched against the column names.
105 :    
106 : parrello 1.1 =item warn
107 :    
108 :     Create an event in the RSS feed when an error occurs.
109 :    
110 :     =item phone
111 :    
112 :     Phone number to message when the script is complete.
113 :    
114 : parrello 1.6 =item DBD
115 :    
116 :     Name of the DBD file. If specified, the DBD must be in the main FIG directory
117 :     (specified in C<$FIG_Config::fig>). This option allows the use of an alternate
118 :     DBD during load, so that access to the database by other processes is not
119 :     compromised.
120 :    
121 : parrello 1.8 =item loadDirectory
122 :    
123 :     Directoty containing the load files. This option allows you to request that
124 :     load files from another version of the NMPDR be used, which is useful when
125 :     creating a new NMPDR: we can yank in the data from the previous database while
126 :     waiting for the new load files to be generated.
127 :    
128 : parrello 1.1 =back
129 :    
130 :     =cut
131 :    
132 :     # Get the command-line options and parameters.
133 :     my ($options, @parameters) = StandardSetup([qw(ERDBLoadGroup ERDB Stats) ],
134 :     {
135 : parrello 1.7 clear => ["", "overwrite existing load files if sections are present"],
136 : parrello 1.4 sanityCheck => ["", "don't load, trace contents of first N load file records instead"],
137 : parrello 1.6 trace => ["2", "tracing level"],
138 : parrello 1.3 keepSections => ["", "if specified, section files will not be deleted after being collated"],
139 : parrello 1.6 phone => ["", "phone number (international format) to call when load finishes"],
140 :     DBD => ["", "if specified, the name of a DBD file in the FIG directory"],
141 : parrello 1.8 loadDirectory => ["", "if specified, an alternate directory containing the load files"],
142 : parrello 1.1 },
143 :     "<database> <group1> <group2> ...",
144 :     @ARGV);
145 :     # Set a variable to contain return type information.
146 :     my $rtype;
147 :     # Insure we catch errors.
148 :     eval {
149 :     # Get the parameters.
150 :     my ($database, @groups) = @parameters;
151 : parrello 1.6 # Check for an alternate DBD.
152 :     my $altDBD = ($options->{DBD} ? "$FIG_Config::fig/$options->{DBD}" : undef);
153 :     # Connect to the database and get its load directory.
154 :     my $erdb = ERDB::GetDatabase($database, $altDBD);
155 : parrello 1.1 # Fix the group list.
156 : parrello 1.3 my @realGroups = ERDBLoadGroup::ComputeGroups($erdb, \@groups);
157 : parrello 1.1 # Get the source object and load directory for this database.
158 :     my $source = $erdb->GetSourceObject();
159 : parrello 1.8
160 :     my $directory = $options->{loadDirectory} || $erdb->LoadDirectory();
161 : parrello 1.1 # Get the list of sections.
162 :     my @sectionList = $erdb->SectionList($source);
163 :     # Create a statistics object to track our progress.
164 :     my $stats = Stats->new();
165 : parrello 1.4 # Find out if we're doing a sanity check.
166 :     my $sanityCheck = $options->{sanityCheck} || "";
167 : parrello 1.2 # Start a timer.
168 :     my $totalStart = time();
169 : parrello 1.1 # Loop through the groups.
170 :     for my $group (@realGroups) {
171 :     # Get the list of tables for this group.
172 : parrello 1.3 my @tableList = ERDBLoadGroup::GetTables($erdb, $group);
173 : parrello 1.2 # We need to insure there is a data file for every table. If we fail to find one,
174 :     # we set the following error flag, which prevents us from loading the database.
175 :     my $missingTable = 0;
176 :     # Loop through the tables in this group.
177 : parrello 1.3 for my $table (@tableList) {
178 : parrello 1.4 Trace("Processing table $table for assembly.") if T(2);
179 :     # Get the section file names.
180 :     my @sectionFiles =
181 :     map { ERDBGenerate::CreateFileName($table, $_, 'data', $directory) } @sectionList;
182 : parrello 1.2 # Get the data file name.
183 :     my $dataFile = ERDBGenerate::CreateFileName($table, undef, 'data', $directory);
184 :     # Do we have it?
185 : parrello 1.7 if (-f $dataFile && ! $options->{clear}) {
186 : parrello 1.2 # Yes. This is good news.
187 :     $stats->Add('tables-found' => 1);
188 : parrello 1.4 Trace("Table file found for $table.") if T(3);
189 : parrello 1.2 } else {
190 :     # No, we must build it. Verify that we have all the sections.
191 :     my @missingFiles = grep { ! -f $_ } @sectionFiles;
192 :     # Did we find everything?
193 :     if (scalar @missingFiles) {
194 :     # No! Denote that we have a missing table.
195 : parrello 1.3 $missingTable++;
196 : parrello 1.2 $stats->Add('tables-skipped' => 1);
197 : parrello 1.6 # If the user wants a sanity check, we want to give him some
198 :     # data anyway.
199 :     if ($sanityCheck) {
200 :     # Get some data lines in the sections. Note we stop when we've exceeded
201 :     # the number of lines expected by the sanity check.
202 :     my @lines;
203 :     for my $sectionFile (@sectionFiles) {
204 :     if (-s $sectionFile && scalar(@lines) < $sanityCheck) {
205 :     Trace("Reading from $sectionFile for $table.") if T(3);
206 :     push @lines, Tracer::GetFile($sectionFile);
207 :     }
208 :     }
209 :     # Create a new temporary file.
210 :     my $tmpFile = "$FIG_Config::temp/$table$$.temp.dtx";
211 :     my $oh = Open(undef, ">$tmpFile");
212 :     # Put all the data into it.
213 :     Trace(scalar(@lines) . " data lines found.") if T(3);
214 :     print $oh join("\n", @lines);
215 :     close $oh;
216 :     # Sanity check the temp file.
217 :     CheckLoadFile($erdb, $table, $tmpFile, $sanityCheck);
218 :     # Clean it up.
219 :     unlink $tmpFile;
220 :     } else {
221 :     # Otherwise tell the user about all the missing files.
222 :     for my $missingFile (@missingFiles) {
223 :     $stats->Add('sections-missing' => 1);
224 :     $stats->AddMessage("Data file $missingFile not found for table $table.");
225 :     }
226 :     }
227 : parrello 1.2 } else {
228 : parrello 1.6 # We have all the sections. Try to assemble them into a data file.
229 : parrello 1.2 my $sortStart = time();
230 : parrello 1.4 my $sortCommand = $erdb->SortNeeded($table) . " >$dataFile";
231 :     Trace("Sort command: $sortCommand") if T(3);
232 : parrello 1.6 # Pipe to the sort command. Note that we turn on autoflush
233 :     # so there's no buffering.
234 : parrello 1.4 my $oh = Open(undef, "| $sortCommand");
235 : parrello 1.6 select $oh; $| = 1; select STDOUT;
236 :     # Loop through the sections.
237 : parrello 1.2 for my $sectionFile (@sectionFiles) {
238 : parrello 1.4 Trace("Collating $sectionFile.") if T(3);
239 :     $stats->Add("$table-sections" => 1);
240 : parrello 1.6 # Loop through the section file.
241 :     my $ih = Open(undef, "<$sectionFile");
242 :     while (defined (my $line = <$ih>)) {
243 :     print $oh $line;
244 : parrello 1.4 $stats->Add("$table-collations" => 1);
245 : parrello 1.2 }
246 :     }
247 :     # Finish the sort step.
248 : parrello 1.4 Trace("Finishing collate for $table.") if T(3);
249 : parrello 1.2 close $oh;
250 :     $stats->Add('tables-collated' => 1);
251 : parrello 1.4 $stats->Add('collate-time' => time() - $sortStart);
252 :     }
253 :     }
254 :     # Now that we know we have a full data file, we can delete the
255 :     # section files to make room in the data directory. The user can
256 :     # turn this behavior off with the keepSections option.
257 :     if (! $options->{keepSections}) {
258 :     for my $sectionFile (@sectionFiles) {
259 :     if (-e $sectionFile) {
260 :     unlink $sectionFile;
261 :     $stats->Add('files-deleted' => 1);
262 : parrello 1.2 }
263 :     }
264 : parrello 1.4 Trace("Section files for $table deleted.") if T(3);
265 : parrello 1.2 }
266 :     }
267 :     # Were any tables missing?
268 :     if ($missingTable) {
269 :     # Yes, skip this group.
270 :     $stats->Add('groups-skipped' => 1);
271 : parrello 1.6 Trace("Skipping $group group: $missingTable missing tables.") if T(2);
272 : parrello 1.2 } else {
273 : parrello 1.4 # No! Process this group's files.
274 :     if ($sanityCheck eq "") {
275 :     Trace("Loading group $group into database.") if T(2);
276 :     } else {
277 :     Trace("Sanity check for group $group.") if T(2);
278 :     }
279 : parrello 1.2 my $loadStart = time();
280 : parrello 1.3 for my $table (@tableList) {
281 : parrello 1.2 my $fileName = ERDBGenerate::CreateFileName($table, undef, 'data', $directory);
282 : parrello 1.4 # Do we want a real load or a sanity check?
283 :     if ($sanityCheck eq "") {
284 :     # Real load.
285 :     my $newStats = $erdb->LoadTable($fileName, $table, truncate => 1, failOnError => 1);
286 :     $stats->Accumulate($newStats);
287 :     Trace("$fileName loaded into $table.") if T(3);
288 :     } elsif ($sanityCheck > 0) {
289 :     # Here we want a sanity check. Note that if the check value is 0,
290 :     # we don't bother. The user just wants to suppress the load step.
291 :     CheckLoadFile($erdb, $table, $fileName, $sanityCheck);
292 :     }
293 : parrello 1.2 }
294 :     $stats->Add("groups-loaded" => 1);
295 :     $stats->Add('load-time' => 1);
296 :     }
297 : parrello 1.1 }
298 : parrello 1.2 $stats->Add('total-time' => time() - $totalStart);
299 :     # Display the statistics from this run.
300 :     Trace("Statistics for load:\n" . $stats->Show()) if T(2);
301 : parrello 1.1 };
302 :     if ($@) {
303 :     Trace("Script failed with error: $@") if T(0);
304 :     } else {
305 :     Trace("Script complete.") if T(2);
306 :     }
307 :     if ($options->{phone}) {
308 : parrello 1.2 my $msgID = Tracer::SendSMS($options->{phone}, "ERDBLoader completed.");
309 : parrello 1.1 if ($msgID) {
310 :     Trace("Phone message sent with ID $msgID.") if T(2);
311 :     } else {
312 :     Trace("Phone message not sent.") if T(2);
313 :     }
314 :     }
315 :    
316 : parrello 1.8 =head2 Internal Methods
317 :    
318 : parrello 1.4 =head3 CheckLoadFile
319 :    
320 :     CheckLoadFile($erdb, $table, $fileName, $count);
321 :    
322 :     Read the first few records of a load file and trace the contents at level
323 :     2. This allows the user to visually compare the load file contents with
324 :     the database definition.
325 :    
326 :     =over 4
327 :    
328 :     =item erdb
329 :    
330 :     [[ErdbPm]] object describing the database.
331 :    
332 :     =item table
333 :    
334 :     Name of the table to check.
335 :    
336 :     =item fileName
337 :    
338 :     Name of the load file to check.
339 :    
340 :     =item count
341 :    
342 :     Number of records to check.
343 :    
344 :     =back
345 :    
346 :     =cut
347 :    
348 :     sub CheckLoadFile {
349 :     # Get the parameters.
350 :     my ($erdb, $table, $fileName, $count) = @_;
351 :     # Open the file for input.
352 :     my $ih = Open(undef, "<$fileName");
353 :     # Slurp the first N records.
354 :     my @records;
355 :     while (! eof $ih && scalar(@records) < $count) {
356 :     push @records, [ Tracer::GetLine($ih) ];
357 :     }
358 :     my $found = scalar(@records);
359 : parrello 1.6 Trace("$found records for $table found in sanity check using $fileName.") if T(2);
360 : parrello 1.4 # Do we have any data at all?
361 :     if ($found) {
362 :     # Yes. Get the table's descriptor. We use this to determine the field names.
363 :     my $relationData = $erdb->FindRelation($table);
364 :     Confess("Relation $table not found in database.") if (! defined $relationData);
365 :     my @fields = @{$relationData->{Fields}};
366 : parrello 1.6 # If this is a relationship, we need the FROM and TO data.
367 :     my %ends; ($ends{from}, $ends{to}) = $erdb->GetRelationshipEntities($table);
368 : parrello 1.4 # Loop through the fields. We generate one message per field.
369 :     for (my $i = 0; $i <= $#fields; $i++) {
370 :     # Get this field's information.
371 :     my $fieldInfo = $fields[$i];
372 :     my $type = $fieldInfo->{type};
373 : parrello 1.6 my $name = $fieldInfo->{name};
374 :     if ($name =~ /^(from|to)-link$/) {
375 :     # Here it's a relationship link, so add the name of the target table to
376 :     # the type.
377 :     $type .= " ($ends{$1})";
378 :     }
379 : parrello 1.4 # This is going to be a multi-line trace message. We start with the field name and type.
380 : parrello 1.6 my @lines = ("Values for $table($name), type $type:\n");
381 : parrello 1.4 # Loop through the records. We generate one line of data per record.
382 :     for (my $j = 0; $j < $found; $j++) {
383 :     # Get the field value.
384 :     my $field = $records[$j]->[$i];
385 :     # Compute the record label.
386 :     my $line = "Record $j";
387 :     # Check for unusual cases.
388 : parrello 1.6 if (! defined $field || $field eq '') {
389 : parrello 1.4 $line .= "= <empty>";
390 :     } else {
391 :     # Make sure we don't trace something ungodly.
392 :     my $excess = (length $field) - 40;
393 :     if ($excess > 0) {
394 :     $field = substr($field, 0, 40) . " >> + $excess characters";
395 :     }
396 :     $line .= ": $field";
397 :     }
398 :     # Save this line. We indent a little for readability.
399 :     push @lines, " $line";
400 :     }
401 :     # Trace this field.
402 :     Trace(join("\n", @lines)) if T(2);
403 :     }
404 :     }
405 :     }
406 :    
407 :    
408 :     1;

MCS Webmaster
ViewVC Help
Powered by ViewVC 1.0.3