Whitelisting of CPR numbers now works.

os2datascanner · Jan 28, 2016 · cc55096 · cc55096
1 parent f8d5fd5
commit cc55096
Show file tree

Hide file tree

Showing 5 changed files with 10 additions and 8 deletions.
diff --git a/.gitignore b/.gitignore
@@ -8,3 +8,4 @@ django-os2webscanner/django.os2webscanner.egg-info
 *.egg-info
 /webscanner_site/webscanner/local_settings.py
 /scrapy-webscanner/scanner/local_settings.py
+.ropeproject
diff --git a/django-os2webscanner/os2webscanner/templates/os2webscanner/report.html b/django-os2webscanner/os2webscanner/templates/os2webscanner/report.html
@@ -155,7 +155,7 @@ <h2>Resultater</h2>
                     Side
                   </small>
                 </strong>
-              </td>a
+              </td>
               {% endif %}
     </tr>
           </thead>

diff --git a/scrapy-webscanner/scanner/rules/cpr.py b/scrapy-webscanner/scanner/rules/cpr.py
@@ -50,7 +50,8 @@ def __init__(self, do_modulus11, ignore_irrelevant, whitelist=None):
     def execute(self, text):
         """Execute the CPR rule."""
         matches = match_cprs(text, do_modulus11=self.do_modulus11,
-                             ignore_irrelevant=self.ignore_irrelevant)
+                             ignore_irrelevant=self.ignore_irrelevant,
+                             whitelist=self.whitelist)
         return matches
 
 # TODO: Improve
@@ -167,7 +168,7 @@ def modulus11_check(cpr):
 
 
 def match_cprs(text, do_modulus11=True, ignore_irrelevant=True,
-               mask_digits=True):
+               mask_digits=True, whitelist=[]):
     """Return MatchItem objects for each CPR matched in the given text.
 
     If mask_digits is False, then the matches will contain full CPR numbers.

diff --git a/scrapy-webscanner/scanner/scanner/scanner.py b/scrapy-webscanner/scanner/scanner/scanner.py
@@ -42,10 +42,10 @@ def _load_rules(self):
         """Load rules based on Scanner settings."""
         rules = []
         if self.scan_object.do_cpr_scan:
-            rules.append(
-                CPRRule(
-                    do_modulus11=self.scan_object.do_cpr_modulus11,
-                    ignore_irrelevant=self.scan_object.do_cpr_ignore_irrelevant
+            rules.append(CPRRule(
+                do_modulus11=self.scan_object.do_cpr_modulus11,
+                ignore_irrelevant=self.scan_object.do_cpr_ignore_irrelevant,
+                whitelist=self.scan_object.whitelisted_cprs
                 )
             )
         if self.scan_object.do_name_scan:

diff --git a/sql/updates/20160127.sql b/sql/updates/20160127.sql
@@ -1,2 +1,2 @@
 ALTER TABLE os2webscanner_organization ADD "cpr_whitelist" text NOT NULL DEFAULT '';
-ALTER TABLE os2webscanner_scanner ADD  "whitelisted_cprs" text NOT NULL DEFAULT '';
+ALTER TABLE os2webscanner_scan ADD  "whitelisted_cprs" text NOT NULL DEFAULT '';