fix: support Created N other repositories record in full activity log

CMUSTRUDEL · Jan 23, 2019 · fc5df30 · fc5df30
1 parent 8b9f537
commit fc5df30
Show file tree

Hide file tree

Showing 3 changed files with 43 additions and 10 deletions.
diff --git a/fixtures/record/created_other_repositories.html b/fixtures/record/created_other_repositories.html
@@ -0,0 +1,34 @@
+<div class="profile-rollup-wrapper py-4 pl-4 position-relative ml-3 js-details-container Details open">
+<span class="discussion-item-icon"><svg aria-hidden="true" class="octicon octicon-repo" height="16" version="1.1" viewbox="0 0 12 16" width="12"><path d="M4 9H3V8h1v1zm0-3H3v1h1V6zm0-2H3v1h1V4zm0-2H3v1h1V2zm8-1v12c0 .55-.45 1-1 1H6v2l-1.5-1.5L3 16v-2H1c-.55 0-1-.45-1-1V1c0-.55.45-1 1-1h10c.55 0 1 .45 1 1zm-1 10H1v2h2v-1h3v1h5v-2zm0-10H2v9h9V1z" fill-rule="evenodd"></path></svg></span>
+<button aria-expanded="false" class="btn-link f4 muted-link no-underline lh-condensed width-full js-details-target" type="button">
+<span class="float-left">
+        Created
+          2
+        other
+        repositories
+      </span>
+<span class="d-inline-block float-right">
+<span aria-label="Collapse" class=" profile-rollup-toggle-closed float-right">
+<svg aria-hidden="true" class="octicon octicon-fold" height="16" version="1.1" viewbox="0 0 14 16" width="14"><path d="M7 9l3 3H8v3H6v-3H4l3-3zm3-6H8V0H6v3H4l3 3 3-3zm4 2c0-.55-.45-1-1-1h-2.5l-1 1h3l-2 2h-7l-2-2h3l-1-1H1c-.55 0-1 .45-1 1l2.5 2.5L0 10c0 .55.45 1 1 1h2.5l1-1h-3l2-2h7l2 2h-3l1 1H13c.55 0 1-.45 1-1l-2.5-2.5L14 5z" fill-rule="evenodd"></path></svg>
+</span>
+<span aria-label="Expand" class="profile-rollup-toggle-open float-right">
+<svg aria-hidden="true" class="octicon octicon-unfold" height="16" version="1.1" viewbox="0 0 14 16" width="14"><path d="M11.5 7.5L14 10c0 .55-.45 1-1 1H9v-1h3.5l-2-2h-7l-2 2H5v1H1c-.55 0-1-.45-1-1l2.5-2.5L0 5c0-.55.45-1 1-1h4v1H1.5l2 2h7l2-2H9V4h4c.55 0 1 .45 1 1l-2.5 2.5zM6 6h2V3h2L7 0 4 3h2v3zm2 3H6v3H4l3 3 3-3H8V9z" fill-rule="evenodd"></path></svg>
+</span>
+</span>
+</button>
+<ul class="profile-rollup-content mt-1" data-repository-hovercards-enabled="">
+<li class="d-block mt-1 py-1" id="contribution-created-repository-1557464-2011-04-01">
+<span class="css-truncate css-truncate-target">
+<span class="profile-rollup-icon">
+<svg aria-hidden="true" class="octicon octicon-repo v-align-middle text-gray-light mr-1" height="16" version="1.1" viewbox="0 0 12 16" width="12"><path d="M4 9H3V8h1v1zm0-3H3v1h1V6zm0-2H3v1h1V4zm0-2H3v1h1V2zm8-1v12c0 .55-.45 1-1 1H6v2l-1.5-1.5L3 16v-2H1c-.55 0-1-.45-1-1V1c0-.55.45-1 1-1h10c.55 0 1 .45 1 1zm-1 10H1v2h2v-1h3v1h5v-2zm0-10H2v9h9V1z" fill-rule="evenodd"></path></svg>
+</span>
+<a class="mr-2" data-hovercard-type="repository" data-hovercard-url="/Vanuan/npapi-cpp-sdk/hovercard" href="/Vanuan/npapi-cpp-sdk">Vanuan/npapi-cpp-sdk</a>
+</span>
+<span class="f6 text-gray">
+</span>
+<time class="float-right f6 text-gray-light pt-1" title="This contribution was made on Apr 1">
+            Apr 1
+          </time>
+</li>
+</ul>
+</div>
diff --git a/fixtures/record/record_test.csv b/fixtures/record/record_test.csv
@@ -6,6 +6,7 @@
 "created_issues.html","{""user2589/Q"": {""issues"": 13}}"
 "created_repositories.html","{""vmarkovtsev/go-lcss"": {""created_repository"": 1}, ""vmarkovtsev/documentation"": {""created_repository"": 1}, ""vmarkovtsev/NMLM"": {""created_repository"": 1}, ""vmarkovtsev/pyupgrade-opt"": {""created_repository"": 1}, ""vmarkovtsev/etalpmet"": {""created_repository"": 1}, ""vmarkovtsev/rgf"": {""created_repository"": 1}}"
 "created_repository.html","{""user2589/Toggl.py"": {""created_repository"": 1}}"
+"created_other_repositories.html","{""Vanuan/npapi-cpp-sdk"": {""created_repository"": 1}}"
 "joined_github.html","{}"
 "created_pull_requests.html","{""mgdg25/Sandbox"": {""pull_requests"": 1}}"
 "created_other_pull_requests.html","{""bblfsh/documentation"": {""pull_requests"": 1}, ""nantes-machine-learning-meetup/NMLM"": {""pull_requests"": 1}, ""spacetelescope/asdf"": {""pull_requests"": 4}, ""asottile/pyupgrade"": {""pull_requests"": 3}, ""src-d/style-analyzer"": {""pull_requests"": 10}, ""src-d/wmd-relax"": {""pull_requests"": 2}, ""vmarkovtsev/pyupgrade-opt"": {""pull_requests"": 2}, ""src-d/guide"": {""pull_requests"": 5}, ""src-d/hercules"": {""pull_requests"": 8}, ""RGF-team/rgf"": {""pull_requests"": 1}, ""src-d/lapjv"": {""pull_requests"": 1}, ""bblfsh/client-python"": {""pull_requests"": 1}, ""src-d/conferences"": {""pull_requests"": 1}, ""src-d/blog"": {""pull_requests"": 3}}"

diff --git a/stgithub.py b/stgithub.py
@@ -116,7 +116,7 @@ def _parse_timeline_update_record(record_div):
         # reviewed pull requests
         title = normalize_text(record_div.button.text)
         if re.match(
-                'Reviewed \\d+ pull requests? in \\d+ repositor(y|ies)', title):
+                r'Reviewed \d+ pull requests? in \d+ repositor(y|ies)', title):
 
             for repo_div in record_div.find_all(
                     'div', class_='profile-rollup-summarized'):
@@ -125,7 +125,7 @@ def _parse_timeline_update_record(record_div):
                 count = int(count_span.text.split()[0])
                 record_data[repo]['reviews'] += count
 
-        elif re.match('Opened \\d+ (?:other )?issues? in \\d+ repositor(y|ies)',
+        elif re.match(r'Opened \d+ (?:other )?issues? in \d+ repositor(y|ies)',
                       title):
             for repo_div in record_div.find_all(
                     'div', class_='profile-rollup-summarized'):
@@ -137,13 +137,13 @@ def _parse_timeline_update_record(record_div):
                     count += int(span.text)
                 record_data[repo]['issues'] += count
 
-        elif re.match('Created \\d+ repositor(y|ies)', title):
+        elif re.match(r'Created \d+ (?:other )?repositor(y|ies)', title):
             for link in record_div.find_all(
                     'a', attrs={'data-hovercard-type': "repository"}):
                 record_data[link.text]['created_repository'] = 1
 
-        elif re.match('Opened \\d+ (?:other )?pull requests? '
-                      'in \\d+ repositor(y|ies)', title):
+        elif re.match(r'Opened \d+ (?:other )?pull requests? '
+                      r'in \d+ repositor(y|ies)', title):
             for repo_div in record_div.find_all(
                     'div', class_='profile-rollup-summarized'):
                 repo = repo_div.button.div.span.text
@@ -154,7 +154,7 @@ def _parse_timeline_update_record(record_div):
                     count += int(span.text)
                 record_data[repo]['pull_requests'] += count
 
-        elif re.match('Created \\d+ commits? in \\d+ repositor(y|ies)', title):
+        elif re.match(r'Created \d+ commits? in \d+ repositor(y|ies)', title):
             for repo_li in record_div.ul.find_all('li', recursive=False):
                 li_div = repo_li.div
                 if not li_div:
@@ -302,7 +302,6 @@ class Scraper(object):
         typically takes couple minutes.
         Use this "API" with caution as it might be extremely slow.
 
-    This class provides access to several functions available
 
     """
     _instance = None  # singleton instance
@@ -426,9 +425,8 @@ def links_to_recent_user_activity(self, user):
         Args:
             user (str): The GitHub login of the user.
 
-        Returns:
-            Generator: A generator of two-tuples:
-                (<%Y-%m-%d date>, link to the activity)
+        Yields:
+            Tuple[str, str]: (<%Y-%m-%d date>, link to the activity)
                 It seems like this feed only includes tags and commits
 
         >>> list(Scraper().links_to_recent_user_activity('user2589'))  # doctest: +SKIP