Create a python script to triage issues

* An issue is considered triaged if certain criterion such as labels are set.

See: kubeflow/community#280

notifications/notifications.py

@@ -12,6 +12,9 @@
 TOKEN_NAME = "GITHUB_TOKEN"
 PULL_REQUEST_TYPE = "PullRequest"
 
+# TODO(jlewi): Rewrite this code to use:
+#  i) graphql.unpack_and_split_nodes
+#  ii) graphql.shard_writer
 def process_notification(n):
   # Mark as read anything that isn't an explicit mention.
   # For PRs there doesn't seem like a simple way to detect if the notice

py/code_intelligence/graphql.py

@@ -1,5 +1,6 @@
 """This module contains utilities for working with GitHub's graphql API"""
-
+from code_intelligence import util
+import logging
 import os
 import requests
 
@@ -12,12 +13,71 @@ def __init__(self):
     self._headers = {"Authorization":
                      "Bearer {0}".format(os.getenv("GITHUB_TOKEN"))}
 
-  def run_query(self, query):
-    """Issue the GraphQL query and return the results."""
+  def run_query(self, query, variables=None):
+    """Issue the GraphQL query and return the results.
+
+    Args:
+     query: String containing the query
+     variables: Dictionary of variables
+    """
+    payload = {'query': query}
+
+    if variables:
+      payload["variables"] = variables
+
     request = requests.post('https://api.github.com/graphql',
-                            json={'query': query}, headers=self._headers)
+                            json=payload, headers=self._headers)
     if request.status_code == 200:
       return request.json()
     else:
       raise Exception("Query failed to run by returning code of {}. {}".format(
-        request.status_code, query))
+        request.status_code, query))
+
+def unpack_and_split_nodes(data, path):
+  """Unpack a list of results
+
+  Args:
+    data: A dictionary containing the results
+    path: A list of fields indicating the fields to select.
+      final one should be a list of nodes
+
+  Returns:
+    issues: A list of dicts; each dict is the data for some of
+    the results
+  """
+
+  children = [data]
+
+  for i, f in enumerate(path):
+    last_child = children[-1]
+    if not f in last_child:
+      # If there are no edges then the field will not exist
+      return []
+    children.append(last_child.get(f))
+
+
+  child = children[-1]
+
+  items = []
+  for i in child:
+    items.append(i["node"])
+
+  return items
+
+class ShardWriter(object):
+  """Write items as a set of file shards"""
+
+  def __init__(self, total_shards, output_dir, prefix="items"):
+    self.output_dir = output_dir
+    self.total_shards = total_shards
+    self.shard = 0
+    self.prefix = prefix
+
+  def write_shard(self, items):
+    """Write the shard"""
+    shard_file = os.path.join(
+      self.output_dir,
+      self.prefix + "-{0:03d}-of-{1:03d}.json".format(
+        self.shard, self.total_shards))
+    util.write_items_to_json(shard_file, items)
+    self.shard += 1

py/code_intelligence/test_data/issues_for_triage.json

@@ -0,0 +1,3 @@
+{"author": {"__typename": "User", "login": "jlewi"}, "title": "Doesn't need triage", "body": "", "url": "https://github.com/kubeflow/kubeflow/issues/365", "labels": {"totalCount": 2, "edges": [{"node": {"name": "kind/bug"}}, {"node": {"name": "area/jupyter"}}, {"node": {"name": "priority/p1"}}]}, "projectCards": {"totalCount": 1, "edges": [{"node": {"project": {"name": "0.2 Release", "number": 2}}}]}}
+{"author": {"__typename": "User", "login": "jlewi"}, "title": "Needs triage; missing everything", "body": "", "url": "https://github.com/kubeflow/kubeflow/issues/365", "labels": {"totalCount": 2, "edges": []}, "projectCards": {"totalCount": 1, "edges": []}}
+{"author": {"__typename": "User", "login": "jlewi"}, "title": "Needs triage; missing projectCards", "body": "", "url": "https://github.com/kubeflow/kubeflow/issues/365", "labels": {"totalCount": 2, "edges": []}}