plotextractor: improve TeX detection

* Made a minor change to the TeX detection step, when looking
  for possible LaTeX sources in a tarball, that allows for
  capitalized file extensions.

* Tarball's and PDF's are now only harvested when needed.

modules/miscutil/lib/plotextractor_converter.py

@@ -96,7 +96,7 @@ def untar(original_tarball, sdir):
         # if neither, maybe it is TeX or an image anyway, otherwise,
         # we don't care
         else:
-            if extracted_file.split('.')[-1] == tex_file_extension:
+            if extracted_file.split('.')[-1].lower() == tex_file_extension:
                 # we might have tex source!
                 might_be_tex.append(extracted_file)
             elif extracted_file.split('.')[-1] in ['eps', 'png', \

modules/miscutil/lib/plotextractor_getter.py

@@ -283,7 +283,7 @@ def tarballs_by_arXiv_id(arXiv_ids, sdir):
     for arXiv_id in arXiv_ids:
         if 'arXiv' not in arXiv_id:
             arXiv_id = 'arXiv:' + arXiv_id
-        tarball, dummy_pdf = harvest_single(arXiv_id, sdir)
+        tarball, dummy_pdf = harvest_single(arXiv_id, sdir, ("tarball",))
         if tarball != None:
             tarballs.append(tarball)
 
@@ -454,13 +454,13 @@ def src_pdf_from_marc(marc_file):
     if possible_match != None:
         # it's listed on arXiv, hooray!
         arXiv_id = possible_match.group(0)
-        dummy1, pdf_loc = harvest_single(arXiv_id, to_dir)
+        dummy1, pdf_loc = harvest_single(arXiv_id, to_dir, ("pdf",))
 
     possible_match = re.search(DESY_match, marc_text)
     if possible_match != None:
         # it's listed on DESY, hooray!
         desy_id = possible_match.group(0)
-        dummy1, pdf_loc = harvest_single(desy_id, to_dir)
+        dummy1, pdf_loc = harvest_single(desy_id, to_dir, ("pdf",))
 
     return pdf_loc