Merge pull request #9 from dcppc/ACharbonneau-patch-1

Update config_centillion.json

.gitignore

@@ -1,4 +1,4 @@
-config_*
+config_flask.py
 vp
 credentials.json
 drive*.json

.gitmodules

@@ -1,3 +1,3 @@
-[submodule "mkdocs-material"]
-	path = mkdocs-material
-	url = https://git.charlesreid1.com/charlesreid1/mkdocs-material.git
+[submodule "mkdocs-material-dib"]
+	path = mkdocs-material-dib
+	url = https://github.com/dib-lab/mkdocs-material-dib.git

Readme.md

@@ -1,18 +1,19 @@
 # The Centillion
 
-**the centillion**: a pan-github-markdown-issues-google-docs search engine.
+**centillion**: a pan-github-markdown-issues-google-docs search engine.
 
 **a centillion**: a very large number consisting of a 1 with 303 zeros after it.
 
-the centillion is 3.03 log-times better than the googol.
+one centillion is 3.03 log-times better than a googol.
 
-![Screen shot of centillion](img/ss.png)
+![Screen shot of centillion](docs/images/ss.png)
 
 
 ## what is it
 
-The centillion is a search engine built using [whoosh](https://whoosh.readthedocs.io/en/latest/intro.html),
-a Python library for building search engines.
+Centillion (https://github.com/dcppc/centillion) is a search engine that can index 
+three kinds of collections: Google Documents, Github issues, and Markdown files in 
+Github repos.
 
 We define the types of documents the centillion should index,
 what info and how. The centillion then builds and
@@ -24,6 +25,30 @@ defined in `centillion.py`.
 
 The centillion keeps it simple.
 
+## authentication layer
+
+Centillion lives behind a Github authentication layer, implemented with 
+[flask-dance](https://github.com/singingwolfboy/flask-dance). When you first
+visit the site it will ask you to authenticate with Github so that it can 
+verify you have permission to access the site.
+
+## technologies
+
+Centillion is a Python program built using whoosh (search engine library). It
+indexes the full text of docx files in Google Documents, just the filenames for
+non-docx files. The full text of issues and their comments are indexed, and
+results are grouped by issue. Centillion requires Google Drive and Github OAuth
+apps. Once you provide credentials to Flask you're all set to go. 
+
+
+## control panel
+
+There's also a control panel at <https://search.nihdatacommons.us/control_panel> 
+that allows you to rebuild the search index from scratch (the Google Drive indexing 
+takes a while).
+
+![Screen shot of centillion control panel](docs/images/cp.png)
+
 
 ## quickstart (with Github auth)
 
@@ -31,6 +56,8 @@ Start by creating a Github OAuth application.
 Get the public and private application key 
 (client token and client secret token)
 from the Github application's page.
+You will also need a Github access token
+(in addition to the app tokens).
 
 When you create the application, set the callback
 URL to `/login/github/authorized`, as in:
@@ -65,11 +92,3 @@ as HTTP by Github, even though there is an HTTPS address, and
 everything else seems fine, try deleting the Github OAuth app
 and creating a new one.
 
-
-## more info
-
-For more info see the documentation: <https://charlesreid1.github.io/centillion>
-
-
-
-

centillion.py

@@ -27,10 +27,10 @@ You provide:
 
 
 class UpdateIndexTask(object):
-    def __init__(self, gh_oauth_token, diff_index=False):
+    def __init__(self, gh_access_token, diff_index=False):
         self.diff_index = diff_index
         thread = threading.Thread(target=self.run, args=())
-        self.gh_oauth_token = gh_oauth_token
+        self.gh_access_token = gh_access_token
         thread.daemon = True
         thread.start()
 
@@ -43,8 +43,8 @@ class UpdateIndexTask(object):
         from get_centillion_config import get_centillion_config
         config = get_centillion_config('config_centillion.json')
 
-        search.update_index_markdown(self.gh_oauth_token,config)
-        search.update_index_issues(self.gh_oauth_token,config)
+        search.update_index_issues(self.gh_access_token,config)
+        search.update_index_markdown(self.gh_access_token,config)
         search.update_index_gdocs(config)
 
 
@@ -55,11 +55,11 @@ app.wsgi_app = ProxyFix(app.wsgi_app)
 # Load default config and override config from an environment variable
 app.config.from_pyfile("config_flask.py")
 
-github_bp = make_github_blueprint()
-#github_bp = make_github_blueprint(
-#                        client_id = os.environ.get('GITHUB_OAUTH_CLIENT_ID'),
-#                        client_secret = os.environ.get('GITHUB_OAUTH_CLIENT_SECRET'),
-#                        scope='read:org')
+#github_bp = make_github_blueprint()
+github_bp = make_github_blueprint(
+                        client_id = os.environ.get('GITHUB_OAUTH_CLIENT_ID'),
+                        client_secret = os.environ.get('GITHUB_OAUTH_CLIENT_SECRET'),
+                        scope='read:org')
 
 app.register_blueprint(github_bp, url_prefix="/login")
 
@@ -172,11 +172,13 @@ def update_index():
                 mresp = github.get('/teams/%s/members/%s'%(copper_team_id,username))
                 if mresp.status_code==204:
 
-                    gh_oauth_token = github.token['access_token']
+                    #gh_oauth_token = github.token['access_token']
+                    gh_access_token = app.config['GITHUB_TOKEN']
 
                     # --------------------
                     # Business as usual
-                    UpdateIndexTask(gh_oauth_token, diff_index=False)
+                    UpdateIndexTask(gh_access_token, 
+                                    diff_index=False)
                     flash("Rebuilding index, check console output")
                     return render_template("controlpanel.html", 
                                            totals={})
@@ -216,5 +218,6 @@ def oops(e):
     return contents404
 
 if __name__ == '__main__':
+    os.environ['OAUTHLIB_INSECURE_TRANSPORT'] = 'true'
     app.run(host="0.0.0.0",port=5000)
 

centillion_search.py

@@ -1,7 +1,7 @@
 import shutil
 import html.parser
 
-from github import Github
+from github import Github, GithubException
 import base64
 
 from gdrive_util import GDrive
@@ -252,7 +252,6 @@ class Search:
             with open(fullpath_input, 'wb') as f:
                 f.write(r.content)
 
-
             # Try to convert docx file to plain text
             try:
                 output = pypandoc.convert_file(fullpath_input,
@@ -316,7 +315,7 @@ class Search:
     # to a search index.
 
 
-    def add_issue(self, writer, issue, config, update=True):
+    def add_issue(self, writer, issue, gh_access_token, config, update=True):
         """
         Add a Github issue/comment to a search index.
         """
@@ -368,7 +367,7 @@ class Search:
 
 
 
-    def add_markdown(self, writer, d, config, update=True):
+    def add_markdown(self, writer, d, gh_access_token, config, update=True):
         """
         Use a Github markdown document API record
         to add a markdown document's contents to
@@ -385,18 +384,27 @@ class Search:
         _, fname = os.path.split(fpath)
         _, fext = os.path.splitext(fpath)
 
-        print("Indexing markdown doc %s"%(fname))
+        print("Indexing markdown doc %s from repo %s"%(fname,repo_name))
 
         # Unpack the requests response and decode the content
-        response = requests.get(furl)
-        jresponse = response.json()
-        content = ""
-        try:
-            binary_content = re.sub('\n','',jresponse['content'])
-            content = base64.b64decode(binary_content).decode('utf-8')
+        # 
+        # don't forget the headers for private repos!
+        # useful: https://bit.ly/2LSAflS
 
-        except KeyError:
-            print(" > XXXXXXXX Failed to extract 'content' field. You probably hit the rate limit.")
+        headers = {'Authorization' : 'token %s'%(gh_access_token)}
+
+        response = requests.get(furl, headers=headers)
+        if response.status_code==200:
+            jresponse = response.json()
+            content = ""
+            try:
+                binary_content = re.sub('\n','',jresponse['content'])
+                content = base64.b64decode(binary_content).decode('utf-8')
+            except KeyError:
+                print(" > XXXXXXXX Failed to extract 'content' field. You probably hit the rate limit.")
+
+        else:
+            print(" > XXXXXXXX Failed to reach file URL. There may be a problem with authentication/headers.")
             return 
 
         # Now create the actual search index record
@@ -431,6 +439,10 @@ class Search:
     # Define how to update search index
     # using different kinds of collections
 
+
+    # ------------------------------
+    # Google Drive Files/Documents
+
     def update_index_gdocs(self, 
                            config):
         """
@@ -478,7 +490,7 @@ class Search:
         remote_ids = set()
         full_items = {}
         while True:
-            ps = 12
+            ps = 100
             results = drive.list(
                     pageSize=ps,
                     pageToken=nextPageToken,
@@ -496,11 +508,11 @@ class Search:
                 # Also store the doc
                 full_items[f['id']] = f
             
-            # Shorter:
-            break
-            ## Longer:
-            #if nextPageToken is None:
-            #    break
+            ## Shorter:
+            #break
+            # Longer:
+            if nextPageToken is None:
+                break
 
 
         writer = self.ix.writer()
@@ -544,13 +556,13 @@ class Search:
         print("Done, updated %d documents in the index" % count)
 
 
+    # ------------------------------
+    # Github Issues/Comments
 
-    def update_index_issues(self, gh_oauth_token, config):
+    def update_index_issues(self, gh_access_token, config):
         """
         Update the search index using a collection of 
         Github repo issues and comments.
-
-        gh_oauth_token can also be an access token.
         """
         # Updated algorithm:
         # - get set of indexed ids
@@ -572,25 +584,29 @@ class Search:
         # Get the set of remote ids:
         # ------
         # Start with api object
-        g = Github(gh_oauth_token)
+        g = Github(gh_access_token)
 
         # Now index all issue threads in the user-specified repos
 
+        # Start by collecting all the things
+        remote_issues = set()
+        full_items = {}
+
         # Iterate over each repo 
         list_of_repos = config['repositories']
         for r in list_of_repos:
 
-            # Start by collecting all the things
-            remote_issues = set()
-            full_items = {}
-
             if '/' not in r:
                 err = "Error: specify org/reponame or user/reponame in list of repos"
                 raise Exception(err)
 
             this_org, this_repo = re.split('/',r)
-            org = g.get_organization(this_org)
-            repo = org.get_repo(this_repo)
+            try:
+                org = g.get_organization(this_org)
+                repo = org.get_repo(this_repo)
+            except:
+                print("Error: could not gain access to repository %s"%(r))
+                continue
 
             # Iterate over each issue thread
             issues = repo.get_issues()
@@ -622,7 +638,7 @@ class Search:
             # cop out
             writer.delete_by_term('id',update_issue)
             item = full_items[update_issue]
-            self.add_issue(writer, item, config, update=True)
+            self.add_issue(writer, item, gh_access_token, config, update=True)
             count += 1
 
 
@@ -631,7 +647,7 @@ class Search:
         add_issues = remote_issues - indexed_issues
         for add_issue in add_issues:
             item = full_items[add_issue]
-            self.add_issue(writer, item, config, update=False)
+            self.add_issue(writer, item, gh_access_token, config, update=False)
             count += 1
 
 
@@ -640,13 +656,13 @@ class Search:
 
 
 
+    # ------------------------------
+    # Github Markdown Files
 
-    def update_index_markdown(self, gh_oauth_token, config): 
+    def update_index_markdown(self, gh_access_token, config): 
         """
         Update the search index using a collection of 
         Markdown files from a Github repo.
-
-        gh_oauth_token can also be an access token.
         """
         EXT = '.md'
 
@@ -669,38 +685,48 @@ class Search:
         # Get the set of remote ids:
         # ------
         # Start with api object
-        g = Github(gh_oauth_token)
+        g = Github(gh_access_token)
 
         # Now index all markdown files
         # in the user-specified repos
 
+        # Start by collecting all the things
+        remote_ids = set()
+        full_items = {}
+
         # Iterate over each repo 
         list_of_repos = config['repositories']
         for r in list_of_repos:
 
-            # Start by collecting all the things
-            remote_ids = set()
-            full_items = {}
-
             if '/' not in r:
                 err = "Error: specify org/reponame or user/reponame in list of repos"
                 raise Exception(err)
 
             this_org, this_repo = re.split('/',r)
-            org = g.get_organization(this_org)
-            repo = org.get_repo(this_repo)
+            try:
+                org = g.get_organization(this_org)
+                repo = org.get_repo(this_repo)
+            except:
+                print("Error: could not gain access to repository %s"%(r))
+                continue
+
 
             # ---------
             # begin markdown-specific code
 
             # Get head commit
             commits = repo.get_commits()
-            last = commits[0]
-            sha = last.sha
+            try:
+                last = commits[0]
+                sha = last.sha
+            except GithubException:
+                print("Error: could not get commits from repository %s"%(r))
+                continue
 
             # Get all the docs
             tree = repo.get_git_tree(sha=sha, recursive=True)
             docs = tree.raw_data['tree']
+            print("Parsing doc ids from repository %s"%(r))
 
             for d in docs:
 
@@ -736,10 +762,10 @@ class Search:
         # and in remote_ids
         update_ids = indexed_ids & remote_ids
         for update_id in update_ids:
-            # cop out
+            # cop out: just delete and re-add
             writer.delete_by_term('id',update_id)
             item = full_items[update_id]
-            self.add_markdown(writer, item, config, update=True)
+            self.add_markdown(writer, item, gh_access_token, config, update=True)
             count += 1
 
 
@@ -748,7 +774,7 @@ class Search:
         add_ids = remote_ids - indexed_ids
         for add_id in add_ids:
             item = full_items[add_id]
-            self.add_markdown(writer, item, config, update=False)
+            self.add_markdown(writer, item, gh_access_token, config, update=False)
             count += 1
 
 
@@ -757,6 +783,16 @@ class Search:
 
 
 
+    # ------------------------------
+    # Groups.io Emails
+
+
+    #def update_index_markdown(self, gh_access_token, config): 
+
+
+
+
+
     # ---------------------------------
     # Search results bundler
 

config_centillion.json

@@ -1,6 +1,27 @@
 {
     "repositories" : [
+        "dcppc/project-management",
+        "dcppc/nih-demo-meetings",
+        "dcppc/internal",
+        "dcppc/organize",
+        "dcppc/dcppc-bot",
+        "dcppc/full-stacks",
+        "dcppc/markdown-issues",
+        "dcppc/design-guidelines-discuss",
+        "dcppc/dcppc-deliverables",
+        "dcppc/dcppc-milestones",
+        "dcppc/crosscut-metadata",
+        "dcppc/lucky-penny",
+        "dcppc/dcppc-workshops",
+        "dcppc/metadata-matrix",
+        "dcppc/data-stewards",
+        "dcppc/dcppc-phase1-demos",
+        "dcppc/apis",
         "dcppc/2018-june-workshop",
-        "dcppc/2018-july-workshop"
+        "dcppc/2018-july-workshop",
+        "dcppc/2018-august-workshop",
+        "dcppc/2018-september-workshop",
+        "dcppc/design-guidelines",
+        "dcppc/2018-may-workshop"
     ]
 }

config_flask.example.py

@@ -2,17 +2,18 @@
 INDEX_DIR = "search_index"
 
 # oauth client deets
-GITHUB_OAUTH_CLIENT_ID = "63f8d49c651840cbe31e"
-GITHUB_OAUTH_CLIENT_SECRET = "36d9a4611f7427336d3c89ed041c45d086b793ee"
+GITHUB_OAUTH_CLIENT_ID = "XXX"
+GITHUB_OAUTH_CLIENT_SECRET = "YYY"
+GITHUB_TOKEN = "ZZZ"
 
 # More information footer: Repository label
-FOOTER_REPO_ORG = "charlesreid1"
+FOOTER_REPO_ORG = "dcppc"
 FOOTER_REPO_NAME = "centillion"
 
 # Toggle to show Whoosh parsed query
 SHOW_PARSED_QUERY=True
 
-TAGLINE = "Search all the things"
+TAGLINE = "Search the Data Commons"
 
 # Flask settings
 DEBUG = True

gdrive_util.py

@@ -29,8 +29,7 @@ class GDrive(object):
     ):
         """
         Set up the Google Drive API instance.
-        Factory method: create it and hand it over.
-        Then we're finished.
+        Factory method: create it here, hand it over in get_service().
         """
         self.credentials_file = credentials_file
         self.client_secret_file = client_secret_file
@@ -40,6 +39,9 @@ class GDrive(object):
         self.store = file.Storage(credentials_file)
 
     def get_service(self):
+        """
+        Return an instance of the Google Drive API service.
+        """
 
         creds = self.store.get()
         if not creds or creds.invalid:

templates/layout.html

@@ -1,5 +1,5 @@
 <!doctype html>
-<title>Markdown Search</title>
+<title>Centillion Search Engine</title>
 <link rel="stylesheet" type="text/css" href="{{ url_for('static', filename='style.css') }}">
 <link rel="stylesheet" type="text/css" href="{{ url_for('static', filename='github-markdown.css') }}">
 <link rel="stylesheet" type="text/css" href="{{ url_for('static', filename='bootstrap.min.css') }}">

templates/search.html

@@ -107,12 +107,18 @@
 
                     <div class="url">
                         {% if e.kind=="gdoc" %}
-                            <b>Google Drive File:</b>
-                            <a href='{{e.url}}'>{{e.title}}</a>
-                            (Owner: {{e.owner_name}}, {{e.owner_email}})
+                            {% if e.mimetype=="document" %}
+                                <b>Google Document:</b>
+                                <a href='{{e.url}}'>{{e.title}}</a>
+                                (Type: {{e.mimetype}}, Owner: {{e.owner_name}}, {{e.owner_email}})
+                            {% else %}
+                                <b>Google Drive:</b>
+                                <a href='{{e.url}}'>{{e.title}}</a>
+                                (Type: {{e.mimetype}}, Owner: {{e.owner_name}}, {{e.owner_email}})
+                            {% endif %}
 
                         {% elif e.kind=="issue" %}
-                            <b>Issue:</b>
+                            <b>Github Issue:</b>
                             <a href='{{e.url}}'>{{e.title}}</a>
                             {% if e.github_user %}
                             opened by <a href='https://github.com/{{e.github_user}}'>@{{e.github_user}}</a>
@@ -121,7 +127,7 @@
                             <b>Repository:</b> <a href='{{e.repo_url}}'>{{e.repo_name}}</a>
 
                         {% elif e.kind=="markdown" %}
-                            <b>Markdown:</b>
+                            <b>Github Markdown:</b>
                             <a href='{{e.url}}'>{{e.title}}</a>
                             <br/>
                             <b>Repository:</b> <a href='{{e.repo_url}}'>{{e.repo_name}}</a>
@@ -131,9 +137,15 @@
 
                         {% endif %}
                         <br />
-                        score: {{'%d'  % e.score}}
+                        Score: {{'%d'  % e.score}}
+                    </div>
+                    <div class="markdown-body">
+                        {% if e.content_highlight %}
+                            {{ e.content_highlight|safe}}
+                        {% else %}
+                        <p>(A preview of this document is not available.)</p>
+                        {% endif %}
                     </div>
-                    <div class="markdown-body">{{ e.content_highlight|safe}}</div>
 
                 </li>
             {% endfor %}