extract language information

1 files changed, 31 insertions, 1 deletions

diff --git a/export.py b/export.py
index f28a45d..3106f07 100644
--- a/export.py
+++ b/export.py
@@ -2,8 +2,17 @@
 import re
 import sys
 
+import numpy
 import pandas as pd
 
+def concat(xss):
+    all = []
+
+    for xs in xss:
+        all.extend(xs)
+
+    return all
+
 class Template(object):
     def __init__(self, name, values):
         self.name = name
@@ -14,6 +23,22 @@ class Template(object):
             self.name,
             '\n'.join(['|' + '%s=%s' % (n, v) for (n, v) in self.values]))
 
+def parse_tags(s):
+    return s.replace('\n', '').split(', ')
+
+def extract_languages(tags):
+    langs = []
+
+    for tag in tags:
+        (a, b) = tag.split('::')
+
+        if a == 'implemented-in':
+            langs.append(b)
+        elif a == 'devel' and b.startswith('lang:'):
+            langs.append(b.split(':')[1])
+
+    return list(set(langs))
+
 def export(pkgs, cps, cpf, name):
     pkg_cps = cps[cps['Upstream-Name'] == name]
     srcpkg_names = list(pkg_cps['_srcpkg'])
@@ -25,6 +50,11 @@ def export(pkgs, cps, cpf, name):
     homepages = list(binpkgs['Homepage'])
     # XXX: maybe choose the one that appears the most?
     homepage = homepages[0] if homepages else ''
+    tags = set(concat(
+        [parse_tags(t) for t in binpkgs['Tag'] if not pd.isnull(t)]))
+    print tags
+    langs = [s.title() for s in extract_languages(tags)]
+    print langs
 
     print Template('Entry', [
         ('Name', srcpkg),
@@ -35,7 +65,7 @@ def export(pkgs, cps, cpf, name):
         # XXX get this information from apt-file
         ('Component programs', ''),
         ('VCS checkout command', ''),
-        ('Computer languages', ''),
+        ('Computer languages', ', '.join(langs)),
         ('Status', ''),
         ('Is GNU', 'No')])