Added option to filter for primary_language and updatate to version 0…

….0.4

README.md

@@ -17,7 +17,7 @@ The package can be installed from [PyPI](https://pypi.org/project/repositories-c
 ```python
 import os
 from datetime import datetime
-from repocollector import GithubRepositoriesCollector
+from repocollector.github import GithubRepositoriesCollector
 
 github_crawler = GithubRepositoriesCollector(
                 access_token=os.getenv('GITHUB_ACCESS_TOKEN'),  # or paste your token
@@ -27,7 +27,8 @@ github_crawler = GithubRepositoriesCollector(
                 min_issues=0,
                 min_releases=0,
                 min_stars=0,
-                min_watchers=0)
+                min_watchers=0,
+                primary_language='language') # e.g., python
 
 for repo in github_crawler.collect_repositories():
     print('id:', repo['id']) # e.g., 123456
@@ -88,17 +89,29 @@ optional arguments:
   --min-watchers MIN_WATCHERS
                         collect repositories with at least <min-watchers>
                         watchers (default: 0)
+  --primary-language LANGUAGE
+                        collect repositories written in this language
   --verbose             show log (default: False)
 ```
 
 
 **Important!** The tool requires a personal access token to access the GraphQL APIs. See how to get one [here](https://github.com/settings/tokens).
 Add ```GITHUB_ACCESS_TOKEN=<paste here your token>``` to the environment variables.
 
-
-### Output
+**Output**
 Running the tool from command-line generates an HTML report accessible at *\<dest\>/report.html*.
 
+**Example**
+The following command search for repositories written in python created between 2014-02-01 and 2014-02-03. 
+The report is saved in the folder /tmp/
+
+```
+repositories-collector 2014-02-01 2014-02-03 /tmp/ --primary-language python
+```
+
+
+
+
 
 ## Contributions
 

repocollector/cli.py

@@ -110,6 +110,13 @@ def get_parser():
                         default=0,
                         help='collect repositories with at least <min-watchers> watchers (default: %(default)s)')
 
+    parser.add_argument('--primary-language',
+                        action='store',
+                        dest='primary_language',
+                        type=str,
+                        default=None,
+                        help='collect repositories written in this language')
+
     parser.add_argument('--verbose',
                         action='store_true',
                         dest='verbose',
@@ -134,7 +141,8 @@ def main():
         min_stars=args.min_stars,
         min_releases=args.min_releases,
         min_watchers=args.min_watchers,
-        min_issues=args.min_issues
+        min_issues=args.min_issues,
+        primary_language=args.primary_language
     )
 
     repositories = list()

repocollector/github.py

@@ -7,7 +7,7 @@
 from datetime import datetime
 
 QUERY = """{ search(query: "is:public stars:>=MIN_STARS mirror:false archived:false created:SINCE..UNTIL 
-pushed:>=PUSHED_AFTER", type: REPOSITORY, first: 50 AFTER) { repositoryCount pageInfo { endCursor startCursor 
+pushed:>=PUSHED_AFTER LANGUAGE:LANGUAGE", type: REPOSITORY, first: 50 AFTER) { repositoryCount pageInfo { endCursor startCursor 
 hasNextPage } edges { node { ... on Repository { databaseId defaultBranchRef { name } owner { login } name url description 
 primaryLanguage { name } stargazers { totalCount } watchers { totalCount } releases { totalCount } issues { 
 totalCount } createdAt pushedAt updatedAt hasIssuesEnabled isArchived isDisabled isMirror isFork object(expression: 
@@ -33,8 +33,8 @@ def __init__(self,
                  min_stars: int = 0,
                  min_releases: int = 0,
                  min_watchers: int = 0,
-                 min_issues: int = 0
-                 ):
+                 min_issues: int = 0,
+                 primary_language: str = None):
         """
         Crawl GitHub to extract repositories
 
@@ -46,6 +46,7 @@ def __init__(self,
         :param min_releases: the minimum number of releases the repositories must have
         :param min_watchers: the minimum number of watchers the repositories must have
         :param min_issues: the minimum number of issues the repositories must have
+        :param primary_language: get repositories written in this language
         """
 
         self._token = access_token
@@ -65,6 +66,11 @@ def __init__(self,
         self.query = re.sub('UNTIL', str(self.until), self.query)
         self.query = re.sub('PUSHED_AFTER', self.pushed_after, self.query)
 
+        if primary_language:
+            self.query = re.sub('LANGUAGE:LANGUAGE', f'language:{primary_language}', self.query)
+        else:
+            self.query = re.sub('LANGUAGE:LANGUAGE', '', self.query)
+
     @property
     def quota(self):
         return self._quota

setup.py

@@ -8,7 +8,7 @@
 with open("README.md", "r") as fh:
     long_description = fh.read()
 
-VERSION = "0.0.3"
+VERSION = "0.0.4"
 
 setup(name='repositories_collector',
       version=VERSION,
@@ -25,12 +25,12 @@
           'console_scripts': ['repositories-collector=repocollector.cli:main'],
       },
       classifiers=[
-          "Development Status :: 3 - Alpha",
+          "Development Status :: 5 - Production/Stable",
           "Intended Audience :: Developers",
           "Programming Language :: Python :: 3.7",
           "License :: OSI Approved :: Apache Software License",
           "Topic :: Software Development :: Libraries :: Python Modules",
           "Operating System :: OS Independent"
       ],
-      insall_requires=requirements
+      install_requires=requirements
 )