From dec173ae33d0dbe4a11ddff01b4773ffe996fd3f Mon Sep 17 00:00:00 2001
From: Ozbolt Menegatti <ozbolt.menegatti@gmail.com>
Date: Thu, 14 Feb 2019 14:33:15 +0100
Subject: [PATCH] Restucturing, now words are parsed right after loading one
 file, not after loading all of them. Should be easilly parallelizable now

---
 wani.py | 22 +++++++++++++++-------
 1 file changed, 15 insertions(+), 7 deletions(-)

diff --git a/wani.py b/wani.py
index 7ce397f..44ba831 100644
--- a/wani.py
+++ b/wani.py
@@ -801,7 +801,7 @@ def load_tei_file(filename, skip_id_check, do_msd_translate, pc_tag, status):
             # strange errors, just skip...
             pass
 
-    return words.values()
+    return list(words.values())
 
 class Writer:
     def __init__(self, args):
@@ -932,10 +932,17 @@ class ColocationIds:
     def set_written(self, key):
         self.data[key][2] = True
 
+    def merge_matches(self, matches, new_matches):
+        for _id, nms in new_matches.items():
+            for nm in nms:
+                matches[_id].append(nm)
+                self.add_match(nm[2])
 
-def match_file(words, structures, colocation_ids, matches=None):
-    if matches is None:
-        matches = {s.id: [] for s in structures}
+        return matches
+
+
+def match_file(words, structures):
+    matches = {s.id: [] for s in structures}
 
     for idx, s in enumerate(structures):
         logging.info("{}/{}: {:7s}".format(idx, len(structures), s.id))
@@ -947,11 +954,11 @@ def match_file(words, structures, colocation_ids, matches=None):
                 colocation_id = [s.id] + list(sorted(colocation_id, key=lambda x:x[0]))
                 colocation_id = tuple(colocation_id)
 
-                colocation_ids.add_match(colocation_id)
                 matches[s.id].append((match, reason, colocation_id))
 
     return matches
 
+
 def main(input_file, structures_file, args):
     writer = Writer(args)
     structures = build_structures(structures_file)
@@ -959,10 +966,11 @@ def main(input_file, structures_file, args):
         logging.debug(str(s))
 
     colocation_ids = ColocationIds()
-    matches = None
+    matches = {s.id: [] for s in structures}
 
     for words in load_files(args):
-        matches = match_file(words, structures, colocation_ids, matches)
+        new_matches = match_file(words, structures)
+        matches = colocation_ids.merge_matches(matches, new_matches)
 
     writer.write_out(matches, structures, colocation_ids)