From bb1838da9a163f4d8ad3f68c18ca7c09ad395766 Mon Sep 17 00:00:00 2001
From: Sebastian Riedel <s@kraih.com>
Date: Tue, 30 Jan 2024 15:57:17 +0100
Subject: [PATCH] Add learning data conversion tool

---
 lib/Cavil/Command/learn.pm | 37 ++++++++++++++++++++++++++++---------
 t/command_learn.t          | 22 ++++++++++++++++++++++
 2 files changed, 50 insertions(+), 9 deletions(-)

diff --git a/lib/Cavil/Command/learn.pm b/lib/Cavil/Command/learn.pm
index 525a43f32d..d36976778f 100644
--- a/lib/Cavil/Command/learn.pm
+++ b/lib/Cavil/Command/learn.pm
@@ -24,11 +24,13 @@ has usage       => sub ($self) { $self->extract_usage };
 
 sub run ($self, @args) {
   getopt \@args,
-    'i|input=s'  => \my $input,
-    'o|output=s' => \my $output,
-    'p|patterns' => \my $patterns;
-  die 'Input or output directory is required' unless (defined $output || defined $input);
+    'c|convert=s' => \my $convert,
+    'i|input=s'   => \my $input,
+    'o|output=s'  => \my $output,
+    'p|patterns'  => \my $patterns;
+  die 'Input or output directory is required' unless (defined $output || defined $input || defined $convert);
 
+  return $self->_convert($convert)          if $convert;
   return $self->_output($output, $patterns) if $output;
   return $self->_input($input);
 }
@@ -42,6 +44,20 @@ sub _classify ($db, $name, $license) {
     $license, $checksum)->rows;
 }
 
+sub _convert ($self, $convert) {
+  my $patterns = $self->app->patterns;
+  my $dir      = path($convert);
+
+  for my $old ($dir->list->each) {
+    my $content  = $old->slurp;
+    my $checksum = $patterns->checksum($content);
+    my $new      = $old->sibling("$checksum.txt");
+    $new->spew($content);
+    $old->remove;
+    say "Converted @{[$old->basename]} to @{[$new->basename]}";
+  }
+}
+
 sub _input ($self, $input) {
   my $db = $self->app->pg->db;
 
@@ -144,12 +160,15 @@ Cavil::Command::learn - Cavil learn command
     script/cavil learn -o ./ml-data
     script/cavil learn -p -o  ./ml-data
     script/cavil learn -i ./ml-data
+    script/cavil learn -c ./text-files
 
   Options:
-    -i, --input <dir>    Import snippet classifications from training data
-    -o, --output <dir>   Export snippets for training machine learning models
-    -p, --patterns       Convert license patterns into snippets and export
-                         those instead
-    -h, --help           Show this summary of available options
+    -c, --convert <dir>   Convert a directory with arbitary text files into
+                          training data (this is a destructive operation)
+    -i, --input <dir>     Import snippet classifications from training data
+    -o, --output <dir>    Export snippets for training machine learning models
+    -p, --patterns        Convert license patterns into snippets and export
+                          those instead
+    -h, --help            Show this summary of available options
 
 =cut
diff --git a/t/command_learn.t b/t/command_learn.t
index 15d6a2823a..1c43376479 100644
--- a/t/command_learn.t
+++ b/t/command_learn.t
@@ -129,6 +129,28 @@ subtest 'Snippets added' => sub {
     }
     like $buffer, qr/Imported 0 snippet classifications/, 'no snippets imported';
   };
+
+  subtest 'Convert arbitrary text files' => sub {
+    my $dir = $tmp->child('convert')->make_path;
+    $dir->child('test.txt')->spew("Hello\nCavil\n");
+    $dir->child('test2')->spew("Hello\nAgain\n");
+    ok -e $dir->child('test.txt'), 'file exists';
+    ok -e $dir->child('test2'),    'file exists';
+
+    my $buffer = '';
+    {
+      open my $handle, '>', \$buffer;
+      local *STDOUT = $handle;
+      $app->start('learn', '-c', "$dir");
+    }
+    like $buffer, qr/Converted test.txt to c512411bea5f292484180fb72e5ea0f9.txt/, 'first file';
+    like $buffer, qr/Converted test2 to 00911bf540aebe36e7c2908760515b25.txt/,    'second file';
+
+    ok !-e $dir->child('test.txt'), 'file no longer exists';
+    ok !-e $dir->child('test2'),    'file no longer exists';
+    is $dir->child('c512411bea5f292484180fb72e5ea0f9.txt')->slurp, "Hello\nCavil\n", 'right content';
+    is $dir->child('00911bf540aebe36e7c2908760515b25.txt')->slurp, "Hello\nAgain\n", 'right content';
+  };
 };
 
 done_testing();