RSFC v0.1.6

CITATION.cff

@@ -34,4 +34,4 @@ keywords:
   - indicators
   - fairness
 license: MIT
-version: 0.1.5
+version: 0.1.6

README.md

@@ -26,6 +26,8 @@ Given a repository URL, RSFC will perform a series of checks based on a list of
 - versioning_standards_use
 - support_issue_tracking
 - has_contribution_guidelines
+- project_is_active
+- software_is_containerized
 
 For more information about these RSQIs, you can check https://github.com/EVERSE-ResearchSoftware/indicators. We have plans to implement all of the RSQIs available in that repository.
 

codemeta.json

@@ -48,6 +48,6 @@
     "operatingSystem": "Linux",
     "programmingLanguage": "Python",
     "relatedLink": "https://github.com/EVERSE-ResearchSoftware/indicators",
-    "version": "0.1.5",
+    "version": "0.1.6",
     "developmentStatus": "wip"
 }

pyproject.toml

@@ -1,6 +1,6 @@
 [tool.poetry]
 name = "rsfc"
-version = "0.1.5"
+version = "0.1.6"
 description = "EVERSE Research Software Fairness Checks"
 authors = [
     "Andres Montero <andres.montero.martin@upm.es>",
@@ -15,7 +15,7 @@ homepage = "https://github.com/oeg-upm/rsfc"
 
 [tool.poetry.dependencies]
 python = ">=3.11,<3.13"
-somef = "0.10.1"
+somef = "0.10.3"
 regex = "2024.11.6"
 requests = "2.32.4"
 anyascii = "0.3.2"
@@ -38,7 +38,7 @@ falcon = "3.1.3"
 fastjsonschema = "2.21.1"
 fonttools = "4.58.4"
 idna = "3.10"
-imbalanced-learn = "0.12.4"
+imbalanced-learn = ">=0.14.1,<0.15.0"
 inflect = "7.5.0"
 iniconfig = "2.1.0"
 jinja2 = "3.1.6"
@@ -48,7 +48,7 @@ jsonschema = "4.24.0"
 jsonschema-specifications = "2025.4.1"
 jupyter-core = "5.8.1"
 kiwisolver = "1.4.8"
-lxml = "5.1.0"
+lxml = "^6.1.0"
 markdown = "3.8.1"
 markupsafe = "3.0.2"
 matplotlib = "3.10.3"

requirements.txt

@@ -71,7 +71,7 @@ scikit-learn==1.5.0
 scipy==1.15.3
 six==1.17.0
 snowballstemmer==3.0.1
-somef==0.10.1
+somef==0.10.3
 soupsieve==2.7
 tabulate==0.9.0
 textblob==0.17.1

src/rsfc/harvesters/github_harvester.py

@@ -1,5 +1,5 @@
 import requests
-from datetime import datetime
+from datetime import datetime, timedelta, timezone
 import urllib
 import yaml
 from rsfc.utils import constants
@@ -19,6 +19,7 @@ def __init__(self, repo_url, branch, tag, token):
         self.codemeta = self.get_codemeta_file()
         self.commits = self.get_commits()
         self.issues = self.get_issues()
+        self.bug_issues = self.get_bugs()
         self.tests = self.get_tests()
 
 
@@ -153,13 +154,16 @@ def get_soft_version(self):
 
 
     def get_commits(self):
+        since = (datetime.now(timezone.utc) - timedelta(days=90)).isoformat()
+        commits_url = ""
+
         if self.repo_type == "GITHUB":
-            commits_url = f"{self.api_url}/commits?sha={self.repo_branch}&per_page=100"
+            commits_url = f"{self.api_url}/commits?sha={self.repo_branch}&since={since}&per_page=100"
             headers = {'Accept': 'application/vnd.github.v3.raw'}
             response = self.safe_request("GET", commits_url, headers=headers)
 
         elif self.repo_type == "GITLAB":
-            commits_url = f"{self.api_url}/repository/commits?ref_name={self.repo_branch}&per_page=100"
+            commits_url = f"{self.api_url}/repository/commits?ref_name={self.repo_branch}&since={since}&per_page=100"
             response = self.safe_request("GET", commits_url)
 
         else:
@@ -171,31 +175,64 @@ def get_commits(self):
             print(f"Error getting commits: {response.status_code}")
             commits = []
 
-        return commits
+        return commits_url, commits
 
 
 
     def get_issues(self):
+        since = (datetime.now(timezone.utc) - timedelta(days=90)).isoformat()
+
         if self.repo_type == "GITHUB":
-            issues_url = f"{self.api_url}/issues?state=all&per_page=100"
+            issues_url = f"{self.api_url}/issues?state=all&since={since}&per_page=100"
             headers = {'Accept': 'application/vnd.github.v3.raw'}
             response = self.safe_request("GET", issues_url, headers=headers)
 
         elif self.repo_type == "GITLAB":
-            issues_url = f"{self.api_url}/issues?state=all&per_page=100"
+            issues_url = f"{self.api_url}/issues?state=all&updated_after={since}&per_page=100"
             response = self.safe_request("GET", issues_url)
 
         else:
             raise ValueError(f"Not supported repository: {self.repo_type}")
 
         issues = []
+
         if response.status_code == 200:
             data = response.json()
-            issues = [issue for issue in data if "pullsafe_request" not in issue]
+            if self.repo_type == "GITHUB":
+                issues = [issue for issue in data if "pull_request" not in issue]
+            else:
+                issues = data
         else:
             print(f"Error getting issues: {response.status_code}")
 
         return issues
+
+    def get_bugs(self):
+
+        if self.repo_type == "GITHUB":
+            issues_url = f"{self.api_url}/issues?state=all&labels=bug&per_page=100"
+            headers = {'Accept': 'application/vnd.github.v3+json'}
+            response = self.safe_request("GET", issues_url, headers=headers)
+
+        elif self.repo_type == "GITLAB":
+            issues_url = f"{self.api_url}/issues?labels=bug&per_page=100"
+            response = self.safe_request("GET", issues_url)
+
+        else:
+            raise ValueError(f"Not supported repository: {self.repo_type}")
+
+        if response.status_code == 200:
+            bugs = response.json()
+
+            if self.repo_type == "GITHUB":
+                bugs = [issue for issue in bugs if "pull_request" not in issue]
+
+        else:
+            print(f"Error getting bugs: {response.status_code}")
+            bugs = []
+
+        return bugs
+
 
 
     def get_tests(self):

src/rsfc/harvesters/somef_harvester.py

@@ -1,70 +1,83 @@
 import io
-import contextlib
 import json
-from somef import somef_cli
-import subprocess
 import os
+import tempfile
+import contextlib
+import subprocess
+
+from somef.somef_cli import run_cli
+
 
 class SomefHarvester:
-
-    def __init__(self, repo_url, branch, tag, token):
+
+    def __init__(self, repo_url, branch=None, tag=None, token=None):
+
         self.somef_configure(token)
-        self.somef_data = self.somef_assessment(repo_url, branch, tag, 0.8)
-
-        
+
+        self.somef_data = self.somef_assessment(repo_url=repo_url, branch=branch, tag=tag, threshold=0.8)
+
     def somef_configure(self, token):
-        
+
         print("Configuring SOMEF...")
-        
+
         if token:
+
             configure = ["somef", "configure"]
+
             stdin_data = (
-            f"{token}\n" #To deal with the inputs asked by somef configure
-            "\n"
-            "\n"
-            "\n"
-            "\n"
-            "\n"
+                f"{token}\n"
+                "\n"
+                "\n"
+                "\n"
+                "\n"
+                "\n"
             )
+
         else:
+
             configure = ["somef", "configure", "-a"]
             stdin_data = None
 
         try:
-            subprocess.run(
-                configure,
-                input=stdin_data,
-                text=True,
-                check=True,
-                stdout=subprocess.DEVNULL,
-                stderr=subprocess.DEVNULL
-            )
+            subprocess.run(configure, input=stdin_data, text=True, check=True, stdout=subprocess.DEVNULL, stderr=subprocess.DEVNULL)
         except subprocess.CalledProcessError as e:
             raise RuntimeError("SOMEF configuration failed") from e
 
     def somef_assessment(self, repo_url, branch=None, tag=None, threshold=0.8):
-    
+
         print("Extracting repository metadata with SOMEF...")
-
+
+        os.makedirs("./rsfc_output/", exist_ok=True)
+
+        output_json = "./rsfc_output/somef_assessment.json"
+
         somef_kwargs = {
             "threshold": threshold,
             "ignore_classifiers": True,
             "repo_url": repo_url,
-            "readme_only": False
+            "readme_only": False,
+            "output": output_json,
+            "pretty": True
         }
 
         if branch is not None:
             somef_kwargs["branch"] = branch
+
         elif tag is not None:
             somef_kwargs["tag"] = tag
 
-        with contextlib.redirect_stdout(io.StringIO()), contextlib.redirect_stderr(io.StringIO()):
-            repo_data = somef_cli.cli_get_data(**somef_kwargs)
-
-        repo_data = json.loads(json.dumps(repo_data.results))
-
-        '''os.makedirs('./rsfc_output/', exist_ok=True)
-        with open('./rsfc_output/somef_assessment.json', 'w', encoding='utf-8') as f:
-            json.dump(repo_data, f, indent=4, ensure_ascii=False)'''
-
+        with (contextlib.redirect_stdout(io.StringIO()), contextlib.redirect_stderr(io.StringIO())):
+
+            run_cli(**somef_kwargs)
+
+        if not os.path.exists(output_json):
+
+            raise RuntimeError(
+                "SOMEF did not generate the expected JSON output"
+            )
+
+        with open(output_json, "r", encoding="utf-8") as f:
+
+            repo_data = json.load(f)
+
         return repo_data