Omit attachments that are not of the expected filetypes

datamade · Nov 5, 2021 · 6e84626 · 6e84626
1 parent f5078d7
commit 6e84626
Show file tree

Hide file tree

Showing 2 changed files with 31 additions and 6 deletions.
diff --git a/scripts/document_link.py b/scripts/document_link.py
@@ -6,17 +6,15 @@
 
 
 def Link(url):
+
     if url.startswith('https://metro.legistar.com/ViewReport.ashx'):
         return BoardReportLink(url)
 
     return DocumentLink(url)
 
 
-class DocumentLink(object):
-    '''
-    https://metro.legistar1.com/metro/meetings/2021/10/2100_A_Operations%2C_Safety%2C_and_Customer_Experience_Committee_21-10-21_Agenda.pdf
-    http://metro.legistar1.com/metro/attachments/6484cdc7-3c2a-4598-abc5-9d846771158e.pdf
-    '''
+class _Link(object):
+
     def __init__(self, url):
         self.url = url
 
@@ -26,6 +24,20 @@ def content(self):
         response = s.get(self.url, retry_on_404=True)
         return response.content
 
+    @property
+    def filename(self):
+        raise NotImplementedError()
+
+    @property
+    def filetype(self):
+        return os.path.splitext(self.filename)[-1]
+
+
+class DocumentLink(_Link):
+    '''
+    https://metro.legistar1.com/metro/meetings/2021/10/2100_A_Operations%2C_Safety%2C_and_Customer_Experience_Committee_21-10-21_Agenda.pdf
+    http://metro.legistar1.com/metro/attachments/6484cdc7-3c2a-4598-abc5-9d846771158e.pdf
+    '''
     @property
     def filename(self):
         '''
@@ -35,7 +47,7 @@ def filename(self):
         return re.sub(r'%', '', os.path.basename(self.url))
 
 
-class BoardReportLink(DocumentLink):
+class BoardReportLink(_Link):
     '''
     https://metro.legistar.com/ViewReport.ashx?M=R&N=TextL5&GID=557&ID=7936&GUID=LATEST&Title=Board+Report
     '''

diff --git a/scripts/download_attachments.py b/scripts/download_attachments.py
@@ -10,13 +10,26 @@
 
 
 if __name__ == '__main__':
+    ACCEPTED_FILETYPES = (
+        '.pdf',
+        '.xlsx',
+        '.doc',
+        '.docx',
+        '.ppt',
+        '.pptx',
+        '.rtf',
+    )
+
     filenames = []
 
     for attachment_link in json.loads(
         os.environ['attachment_links'].replace('\'', '"')
     ):
         link = Link(attachment_link)
 
+        if link.filetype not in ACCEPTED_FILETYPES:
+            continue
+
         with open(os.path.join('attachments', link.filename), 'wb') as file:
             file.write(link.content)