Draft helper script for intermap tables.

[philipp/winterrodeln/wrpylib.git] / wrpylib / wrmwcache.py

diff --git a/wrpylib/wrmwcache.py b/wrpylib/wrmwcache.py
index 1540dc8fe58181311832d0c3793608de3401df7d..95e0e96805f3d0ba8a5af4ff1e97c23890b372c1 100644 (file)
--- a/wrpylib/wrmwcache.py
+++ b/wrpylib/wrmwcache.py
@@ -6,7 +6,7 @@ from sqlalchemy import schema
 from sqlalchemy.sql import select
 from sqlalchemy.sql.expression import func as sqlfunc
 from osgeo import ogr
-from wrpylib import mwdb, wrmwdb, mwmarkup, wrmwmarkup, wrvalidators
+from wrpylib import mwdb, wrmwdb, wrmwmarkup, wrvalidators
 
 
 class UpdateCacheError(RuntimeError):
@@ -50,7 +50,7 @@ def update_wrsledruncache(connection):
             sledrun.page_id = sledrun_page.page_id
             sledrun.page_title = sledrun_page.page_title
             sledrun.name_url = wrvalidators.sledrun_page_title_to_pretty_url(sledrun_page.page_title)
-            sledrun.under_construction = connection.execute(select([categorylinks], (categorylinks.c.cl_from==sledrun_page.page_id) & (categorylinks.c.cl_to == 'In_Arbeit')).alias('x').count()).fetchone()[0] > 0
+            sledrun.under_construction = connection.execute(select([sqlfunc.count()], (categorylinks.c.cl_from==sledrun_page.page_id) & (categorylinks.c.cl_to == 'In_Arbeit')).alias('x')).fetchone()[0] > 0
             connection.execute(wrsledruncache.insert(sledrun.__dict__))
         except ValueError as e:
             transaction.rollback()
@@ -93,7 +93,7 @@ def update_wrinncache(connection):
             inn = wrmwmarkup.inn_from_gasthausbox(gasthausbox, Inn())
             inn.page_id = inn_page.page_id
             inn.page_title = inn_page.page_title
-            inn.under_construction = connection.execute(select([categorylinks], (categorylinks.c.cl_from==inn_page.page_id) & (categorylinks.c.cl_to == 'In_Arbeit')).alias('x').count()).fetchone()[0] > 0 # It would be better to do this in the query above
+            inn.under_construction = connection.execute(select([sqlfunc.count()], (categorylinks.c.cl_from==inn_page.page_id) & (categorylinks.c.cl_to == 'In_Arbeit')).alias('x')).fetchone()[0] > 0 # It would be better to do this in the query above
             connection.execute(wrinncache.insert(inn.__dict__))
         except ValueError as e:
             transaction.rollback()
@@ -172,11 +172,16 @@ def update_wrmapcache(connection):
     
     # Refill wrmappointcache and wrmappathcache tables
     for sledrun_page in sledrun_pages:
-        try: 
-            start, content, endtag, end = mwmarkup.find_tag(sledrun_page.old_text, 'wrmap')
-            if content is None:
-                continue
-            geojson = wrmwmarkup.parse_wrmap(sledrun_page.old_text[start:end])
+        try:
+            import mwparserfromhell
+            wikicode = mwparserfromhell.parse(sledrun_page.old_text)
+            wrmap_list = wikicode.filter_tags(recursive=False, matches=lambda tag: tag.tag == 'wrmap')
+            if len(wrmap_list) == 0:
+                continue  # not wrmap in page
+            if len(wrmap_list) > 1:
+                raise UpdateCacheError('{} <wrmap ...> entries found in article "{}"'.format(len(wrmap_list), sledrun_page.page_title))
+            wrmap = wrmap_list[0]
+            geojson = wrmwmarkup.parse_wrmap(str(wrmap))
 
             for feature in geojson['features']:
                 properties = feature['properties']