osa_tool 0.2.2__tar.gz → 0.2.3__tar.gz

{osa_tool-0.2.2 → osa_tool-0.2.3}/PKG-INFO

@@ -1,17 +1,18 @@
-Metadata-Version: 2.3
+Metadata-Version: 2.4
 Name: osa_tool
-Version: 0.2.2
+Version: 0.2.3
 Summary: Tool that just makes your open source project better!
-License: BSD-3-Clause
+License-Expression: BSD-3-Clause
+License-File: LICENSE
 Author: ITMO-NSS-team
 Author-email: itmo.nss.team@gmail.com
 Requires-Python: >=3.10,<4.0
-Classifier: License :: OSI Approved :: BSD License
 Classifier: Programming Language :: Python :: 3
 Classifier: Programming Language :: Python :: 3.10
 Classifier: Programming Language :: Python :: 3.11
 Classifier: Programming Language :: Python :: 3.12
 Classifier: Programming Language :: Python :: 3.13
+Classifier: Programming Language :: Python :: 3.14
 Requires-Dist: ProtoLLM (>=0.1.5)
 Requires-Dist: PyYAML (==6.0.2)
 Requires-Dist: Pygments (==2.19.1)
@@ -271,7 +272,7 @@ Docker, ensure that you upload the PDF file to the OSA folder before building th
 /app/OSA/... or just use volume mounting to access the file.
 
 The --generate-workflows option is intended to create customizable CI/CD pipelines for Python repositories. For detailed
-documentation, see the [GitHub Action Workflow Generator README](./osa_tool/github_workflow/README.md).
+documentation, see the [Workflow Generator README](./osa_tool/workflow/README.md).
 
 ### Configuration
 

{osa_tool-0.2.2 → osa_tool-0.2.3}/README.md

@@ -185,7 +185,7 @@ Docker, ensure that you upload the PDF file to the OSA folder before building th
 /app/OSA/... or just use volume mounting to access the file.
 
 The --generate-workflows option is intended to create customizable CI/CD pipelines for Python repositories. For detailed
-documentation, see the [GitHub Action Workflow Generator README](./osa_tool/github_workflow/README.md).
+documentation, see the [Workflow Generator README](./osa_tool/workflow/README.md).
 
 ### Configuration
 

{osa_tool-0.2.2 → osa_tool-0.2.3}/osa_tool/aboutgen/about_generator.py

@@ -1,13 +1,10 @@
 import os
 import re
-import time
 from typing import List
 
-import requests
-
 from osa_tool.aboutgen.prompts_about_config import PromptAboutLoader
-from osa_tool.analytics.metadata import detect_platform, load_data_metadata
 from osa_tool.config.settings import ConfigLoader
+from osa_tool.git_agent.git_agent import GitAgent
 from osa_tool.models.models import ModelHandler, ModelHandlerFactory
 from osa_tool.utils import extract_readme_content, logger, parse_folder_name
 
@@ -26,15 +23,15 @@ HOMEPAGE_KEYS = [
 class AboutGenerator:
     """Generates Git repository About section content."""
 
-    def __init__(self, config_loader: ConfigLoader):
+    def __init__(self, config_loader: ConfigLoader, git_agent: GitAgent):
         self.config = config_loader.config
         self.model_handler: ModelHandler = ModelHandlerFactory.build(self.config)
         self.repo_url = self.config.git.repository
-        self.metadata = load_data_metadata(self.repo_url)
+        self.metadata = git_agent.metadata
         self.base_path = os.path.join(os.getcwd(), parse_folder_name(self.repo_url))
         self.readme_content = extract_readme_content(self.base_path)
-        self.platform = detect_platform(self.repo_url)
         self.prompts = PromptAboutLoader().prompts
+        self.validate_topics = git_agent.validate_topics
 
         self._content: dict | None = None
 
@@ -136,128 +133,12 @@ class AboutGenerator:
             response = self.model_handler.send_request(formatted_prompt)
             topics = [topic.strip().lower().replace(" ", "-") for topic in response.split(",") if topic.strip()]
             logger.debug(f"Generated topics from LLM: {topics}")
-            validated_topics = self._validate_topics(topics)
+            validated_topics = self.validate_topics(topics)
             return list({*existing_topics, *validated_topics})
         except Exception as e:
             logger.error(f"Error generating topics: {e}")
             return []
 
-    def _validate_topics(self, topics: List[str]) -> List[str]:
-        """Validates topics against platform-specific APIs.
-
-        Args:
-            topics (List[str]): List of potential topics to validate
-
-        Returns:
-            List[str]: List of validated topics that exist on platform
-        """
-        if self.platform == "github":
-            return self._validate_github_topics(topics)
-        elif self.platform == "gitlab":
-            return self._validate_gitlab_topics(topics)
-        elif self.platform == "gitverse":
-            return self._validate_gitverse_topics(topics)
-        else:
-            logger.warning(f"Topic validation not implemented for platform: {self.platform}")
-            return topics
-
-    @staticmethod
-    def _validate_github_topics(topics: List[str]) -> List[str]:
-        """Validates topics against GitHub Topics API.
-
-        Args:
-            topics (List[str]): List of potential topics to validate
-
-        Returns:
-            List[str]: List of validated topics that exist on GitHub
-        """
-        logger.info("Validating topics against GitHub Topics API...")
-        min_repo = 5
-        validated_topics = []
-
-        for topic in topics:
-            try:
-                response = requests.get(
-                    f"https://api.github.com/search/topics?q={topic}+repositories:>{min_repo}",
-                    headers={"Accept": "application/vnd.github.v3+json"},
-                )
-
-                if response.status_code == 200:
-                    data = response.json()
-                    if (total := data.get("total_count", 0)) > 0:
-                        if total == 1:
-                            valid_topic = data.get("items")[0].get("name")
-                            logger.debug(f"Applied transformation for topic: '{topic} -> {valid_topic}'")
-                        else:
-                            valid_topic = topic
-                        validated_topics.append(valid_topic)
-                    else:
-                        logger.debug(f"Generated topic '{topic}' is not valid, skipping")
-                elif response.status_code == 403:
-                    logger.warning("Rate limit exceeded, waiting 60 seconds")
-                    time.sleep(60)
-
-                time.sleep(1)
-
-            except Exception as e:
-                logger.error(f"Error validating topic '{topic}': {e}")
-                continue
-
-        logger.info(f"Validated {len(validated_topics)} topics out of {len(topics)}.")
-        return validated_topics
-
-    @staticmethod
-    def _validate_gitlab_topics(topics: List[str]) -> List[str]:
-        """Validates topics against GitLab Topics API.
-
-        Args:
-            topics (List[str]): List of potential topics to validate
-
-        Returns:
-            List[str]: List of validated topics that exist on GitLab
-        """
-        logger.info("Validating topics against GitLab Topics API...")
-        validated_topics = []
-        base_url = "https://gitlab.com/api/v4/topics"
-        headers = {"Accept": "application/json"}
-
-        for topic in topics:
-            try:
-                params = {"search": topic}
-                response = requests.get(base_url, headers=headers, params=params)
-                if response.status_code == 200:
-                    data = response.json()
-                    for entry in data:
-                        if entry.get("name") == topic:
-                            validated_topics.append(topic)
-                            logger.debug(f"Validated GitLab topic: {topic}")
-                            break
-                    else:
-                        logger.debug(f"Topic '{topic}' not found on GitLab, skipping")
-                elif response.status_code == 403:
-                    logger.warning("Rate limit exceeded, waiting 60 seconds")
-                    time.sleep(60)
-                time.sleep(1)
-            except Exception as e:
-                logger.error(f"Error validating topic '{topic}': {e}")
-                continue
-
-        logger.info(f"Validated {len(validated_topics)} topics out of {len(topics)}.")
-        return validated_topics
-
-    @staticmethod
-    def _validate_gitverse_topics(topics: List[str]) -> List[str]:
-        """Validates topics for Gitverse platform.
-
-        Args:
-            topics (List[str]): List of potential topics to validate
-
-        Returns:
-            List[str]: List of topics (no validation performed)
-        """
-        logger.warning("Topic validation is not implemented for Gitverse. Returning original topics list.")
-        return topics
-
     def detect_homepage(self) -> str:
         """
         Detects the homepage URL for a project.

osa_tool-0.2.3/osa_tool/analytics/metadata.py

@@ -0,0 +1,361 @@
+import os
+import re
+from abc import ABC, abstractmethod
+from dataclasses import dataclass
+from datetime import datetime
+
+import requests
+from dotenv import load_dotenv
+from requests import HTTPError
+
+from osa_tool.utils import get_base_repo_url, logger
+
+load_dotenv()
+
+
+@dataclass
+class RepositoryMetadata:
+    """
+    Dataclass to store Git repository metadata.
+    """
+
+    name: str
+    full_name: str
+    owner: str
+    owner_url: str | None
+    description: str | None
+
+    # Repository statistics
+    stars_count: int
+    forks_count: int
+    watchers_count: int
+    open_issues_count: int
+
+    # Repository details
+    default_branch: str
+    created_at: str
+    updated_at: str
+    pushed_at: str
+    size_kb: int
+
+    # Repository URLs
+    clone_url_http: str
+    clone_url_ssh: str
+    contributors_url: str | None
+    languages_url: str
+    issues_url: str | None
+
+    # Programming languages and topics
+    language: str | None
+    languages: list[str]
+    topics: list[str]
+
+    # Additional repository settings
+    has_wiki: bool
+    has_issues: bool
+    has_projects: bool
+    is_private: bool
+    homepage_url: str | None
+
+    # License information
+    license_name: str | None
+    license_url: str | None
+
+
+class MetadataLoader(ABC):
+    """
+    Abstract base class for repository metadata loaders.
+    """
+
+    @classmethod
+    def load_data(cls, repo_url: str) -> RepositoryMetadata:
+        """
+        General method to load repository metadata for a given URL.
+        Calls the platform-specific loader method.
+
+        Args:
+            repo_url (str): The full URL of the repository.
+
+        Returns:
+            RepositoryMetadata: Parsed repository metadata.
+        """
+        try:
+            return cls._load_platform_data(repo_url)
+
+        except HTTPError as http_exc:
+            status_code = getattr(http_exc.response, "status_code", None)
+            logger.error(f"Error while fetching repository metadata: {http_exc}")
+
+            if status_code == 401:
+                logger.error("Authentication failed: please check your Git token (missing or expired).")
+            elif status_code == 404:
+                logger.error("Repository not found: please check the repository URL.")
+            elif status_code == 403:
+                logger.error("Access denied: your token may lack sufficient permissions or you hit a rate limit.")
+            else:
+                logger.error("Unexpected HTTP error occurred while accessing the repository metadata.")
+            raise
+
+        except Exception as exc:
+            logger.error(f"Unexpected error while fetching repository metadata: {exc}")
+            raise
+
+    @classmethod
+    @abstractmethod
+    def _load_platform_data(cls, repo_url: str) -> RepositoryMetadata:
+        """
+        Abstract method to load metadata from a platform-specific API.
+
+        Args:
+            repo_url (str): The full URL of the repository.
+
+        Returns:
+            RepositoryMetadata: Parsed repository metadata.
+        """
+        pass
+
+    @classmethod
+    @abstractmethod
+    def _parse_metadata(cls, repo_data: dict) -> RepositoryMetadata:
+        """
+        Abstract method to parse raw API response dictionary into RepositoryMetadata.
+
+        Args:
+            repo_data (dict): Raw API response data.
+
+        Returns:
+            RepositoryMetadata: Parsed repository metadata.
+        """
+        pass
+
+
+class GitHubMetadataLoader(MetadataLoader):
+    @classmethod
+    def _load_platform_data(cls, repo_url: str) -> RepositoryMetadata:
+        """
+        Load GitHub repository metadata via GitHub API.
+
+        Args:
+            repo_url (str): URL of the GitHub repository.
+
+        Returns:
+            RepositoryMetadata: Parsed metadata object.
+        """
+        base_url = get_base_repo_url(repo_url)
+        headers = {
+            "Authorization": f"token {os.getenv('GIT_TOKEN', os.getenv('GITHUB_TOKEN', ''))}",
+            "Accept": "application/vnd.github.v3+json",
+        }
+        url = f"https://api.github.com/repos/{base_url}"
+        response = requests.get(url=url, headers=headers)
+        response.raise_for_status()
+        data = response.json()
+        logger.info(f"Successfully fetched GitHub metadata for repository: '{base_url}'")
+        return GitHubMetadataLoader._parse_metadata(data)
+
+    @classmethod
+    def _parse_metadata(cls, repo_data: dict) -> RepositoryMetadata:
+        """
+        Parse GitHub API response into RepositoryMetadata.
+
+        Args:
+            repo_data (dict): Raw GitHub API response.
+
+        Returns:
+            RepositoryMetadata: Parsed repository metadata.
+        """
+        languages = repo_data.get("languages", {})
+        license_info = repo_data.get("license", {}) or {}
+        owner_info = repo_data.get("owner", {}) or {}
+
+        return RepositoryMetadata(
+            name=repo_data.get("name", ""),
+            full_name=repo_data.get("full_name", ""),
+            owner=owner_info.get("login", ""),
+            owner_url=owner_info.get("html_url", ""),
+            description=repo_data.get("description", ""),
+            stars_count=repo_data.get("stargazers_count", 0),
+            forks_count=repo_data.get("forks_count", 0),
+            watchers_count=repo_data.get("watchers_count", 0),
+            open_issues_count=repo_data.get("open_issues_count", 0),
+            default_branch=repo_data.get("default_branch", ""),
+            created_at=repo_data.get("created_at", ""),
+            updated_at=repo_data.get("updated_at", ""),
+            pushed_at=repo_data.get("pushed_at", ""),
+            size_kb=repo_data.get("size", 0),
+            clone_url_http=repo_data.get("clone_url", ""),
+            clone_url_ssh=repo_data.get("ssh_url", ""),
+            contributors_url=repo_data.get("contributors_url"),
+            languages_url=repo_data.get("languages_url", ""),
+            issues_url=repo_data.get("issues_url"),
+            language=repo_data.get("language", ""),
+            languages=list(languages.keys()) if languages else [],
+            topics=repo_data.get("topics", []),
+            has_wiki=repo_data.get("has_wiki", False),
+            has_issues=repo_data.get("has_issues", False),
+            has_projects=repo_data.get("has_projects", False),
+            is_private=repo_data.get("private", False),
+            homepage_url=repo_data.get("homepage", ""),
+            license_name=license_info.get("name", ""),
+            license_url=license_info.get("url", ""),
+        )
+
+
+class GitLabMetadataLoader(MetadataLoader):
+    @classmethod
+    def _load_platform_data(cls, repo_url: str) -> RepositoryMetadata:
+        """
+        Load GitLab repository metadata via GitLab API.
+
+        Args:
+            repo_url (str): URL of the GitLab repository.
+
+        Returns:
+            RepositoryMetadata: Parsed metadata object.
+        """
+        base_url = get_base_repo_url(repo_url)
+        gitlab_instance_match = re.match(r"(https?://[^/]*gitlab[^/]*)", repo_url)
+        if not gitlab_instance_match:
+            raise ValueError("Invalid GitLab repository URL")
+        gitlab_instance = gitlab_instance_match.group(1)
+
+        headers = {
+            "Authorization": f"Bearer {os.getenv('GITLAB_TOKEN', os.getenv('GIT_TOKEN'))}",
+            "Content-Type": "application/json",
+        }
+        project_path = base_url.replace("/", "%2F")
+        url = f"{gitlab_instance}/api/v4/projects/{project_path}"
+
+        response = requests.get(url=url, headers=headers)
+        response.raise_for_status()
+        data = response.json()
+        logger.info(f"Successfully fetched GitLab metadata for repository: '{base_url}'")
+        return GitLabMetadataLoader._parse_metadata(data)
+
+    @classmethod
+    def _parse_metadata(cls, repo_data: dict) -> RepositoryMetadata:
+        """
+        Parse GitLab API response into RepositoryMetadata.
+
+        Args:
+            repo_data (dict): Raw GitLab API response.
+
+        Returns:
+            RepositoryMetadata: Parsed repository metadata.
+        """
+        owner_info = repo_data.get("owner", {}) or {}
+        namespace = repo_data.get("namespace", {}) or {}
+
+        created_raw = repo_data.get("created_at", "")
+        if created_raw:
+            created_time = datetime.strptime(created_raw.split(".")[0], "%Y-%m-%dT%H:%M:%S").strftime(
+                "%Y-%m-%dT%H:%M:%SZ"
+            )
+        else:
+            created_time = ""
+
+        return RepositoryMetadata(
+            name=repo_data.get("name", ""),
+            full_name=repo_data.get("path_with_namespace", ""),
+            owner=namespace.get("name", "") or owner_info.get("name", ""),
+            owner_url=namespace.get("web_url", "") or owner_info.get("web_url", ""),
+            description=repo_data.get("description", ""),
+            stars_count=repo_data.get("star_count", 0),
+            forks_count=repo_data.get("forks_count", 0),
+            watchers_count=0,  # GitLab does not have watchers, set to 0
+            open_issues_count=repo_data.get("open_issues_count", 0),
+            default_branch=repo_data.get("default_branch", ""),
+            created_at=created_time,
+            updated_at=repo_data.get("last_activity_at", ""),
+            pushed_at=repo_data.get("last_activity_at", ""),
+            size_kb=repo_data.get("repository_size", 0) // 1024,  # Convert bytes to KB
+            clone_url_http=repo_data.get("http_url_to_repo", ""),
+            clone_url_ssh=repo_data.get("ssh_url_to_repo", ""),
+            contributors_url=f"{repo_data.get('web_url', '')}/contributors" if repo_data.get("web_url") else None,
+            languages_url=f"{repo_data.get('web_url', '')}/languages" if repo_data.get("web_url") else "",
+            issues_url=f"{repo_data.get('web_url', '')}/issues" if repo_data.get("web_url") else None,
+            language="",  # GitLab API does not provide primary language
+            languages=[],
+            topics=repo_data.get("tag_list", []),
+            has_wiki=repo_data.get("wiki_enabled", False),
+            has_issues=repo_data.get("issues_enabled", False),
+            has_projects=True,  # GitLab always has project management features enabled
+            is_private=repo_data.get("visibility", "public") != "public",
+            homepage_url="",
+            license_name="",
+            license_url="",
+        )
+
+
+class GitverseMetadataLoader(MetadataLoader):
+    @classmethod
+    def _load_platform_data(cls, repo_url: str) -> RepositoryMetadata:
+        """
+        Load Gitverse repository metadata via Gitverse API.
+
+        Args:
+            repo_url (str): URL of the Gitverse repository.
+
+        Returns:
+            RepositoryMetadata: Parsed metadata object.
+        """
+        base_url = get_base_repo_url(repo_url)
+        headers = {
+            "Authorization": f"Bearer {os.getenv('GITVERSE_TOKEN', os.getenv('GIT_TOKEN'))}",
+            "Accept": "application/vnd.gitverse.object+json;version=1",
+            "Content-Type": "application/json",
+            "User-Agent": "Mozilla/5.0",
+        }
+        url = f"https://api.gitverse.ru/repos/{base_url}"
+
+        response = requests.get(url=url, headers=headers)
+        response.raise_for_status()
+        data = response.json()
+        logger.info(f"Successfully fetched Gitverse metadata for repository: '{base_url}'")
+        return GitverseMetadataLoader._parse_metadata(data)
+
+    @classmethod
+    def _parse_metadata(cls, repo_data: dict) -> RepositoryMetadata:
+        """
+        Parse Gitverse API response into RepositoryMetadata.
+
+        Args:
+            repo_data (dict): Raw Gitverse API response.
+
+        Returns:
+            RepositoryMetadata: Parsed repository metadata.
+        """
+        owner_info = repo_data.get("owner", {}) or {}
+        license_info = repo_data.get("license", {}) or {}
+
+        return RepositoryMetadata(
+            name=repo_data.get("name", ""),
+            full_name=repo_data.get("full_name", ""),
+            owner=owner_info.get("login", ""),
+            owner_url=owner_info.get("html_url", ""),
+            description=repo_data.get("description", ""),
+            stars_count=repo_data.get("stargazers_count", 0),
+            forks_count=repo_data.get("forks_count", 0),
+            watchers_count=repo_data.get("watchers_count", 0),
+            open_issues_count=repo_data.get("open_issues_count", 0),
+            default_branch=repo_data.get("default_branch", ""),
+            created_at=repo_data.get("created_at", ""),
+            updated_at=repo_data.get("updated_at", ""),
+            pushed_at=repo_data.get("pushed_at", ""),
+            size_kb=repo_data.get("size", 0),
+            clone_url_http=repo_data.get("clone_url", ""),
+            clone_url_ssh=repo_data.get("ssh_url", ""),
+            contributors_url=repo_data.get("contributors_url"),
+            languages_url=repo_data.get("languages_url", ""),
+            issues_url=repo_data.get("issues_url"),
+            language=repo_data.get("language", ""),
+            languages=repo_data.get("languages", []) or [],
+            topics=repo_data.get("topics", []) or [],
+            has_wiki=repo_data.get("has_wiki", False),
+            has_issues=repo_data.get("has_issues", False),
+            has_projects=repo_data.get("has_projects", False),
+            is_private=repo_data.get("private", False),
+            homepage_url=repo_data.get("homepage", ""),
+            license_name=license_info.get("name", ""),
+            license_url=license_info.get("url", ""),
+        )

osa_tool-0.2.3/osa_tool/analytics/prompt_builder.py

@@ -0,0 +1,66 @@
+from enum import Enum
+from typing import Annotated
+
+from pydantic import BaseModel
+
+
+class YesNoPartial(str, Enum):
+    YES = "Yes"
+    NO = "No"
+    PARTIAL = "Partial"
+    UNKNOWN = "Unknown"
+
+
+class RepositoryStructure(BaseModel):
+    compliance: Annotated[str, "Compliance with standard structure"] = "Unknown"
+    missing_files: Annotated[
+        list[str],
+        "List of missing critical files that impact project usability and clarity",
+    ] = []
+    organization: Annotated[
+        str,
+        "Evaluation of the overall organization of directories and files for maintainability and clarity",
+    ] = "Unknown"
+
+
+class ReadmeEvaluation(BaseModel):
+    readme_quality: Annotated[str, "Assessment of the README quality with a brief comment"] = "Unknown"
+    project_description: YesNoPartial = YesNoPartial.UNKNOWN
+    installation: YesNoPartial = YesNoPartial.UNKNOWN
+    usage_examples: YesNoPartial = YesNoPartial.UNKNOWN
+    contribution_guidelines: YesNoPartial = YesNoPartial.UNKNOWN
+    license_specified: YesNoPartial = YesNoPartial.UNKNOWN
+    badges_present: YesNoPartial = YesNoPartial.UNKNOWN
+
+
+class CodeDocumentation(BaseModel):
+    tests_present: YesNoPartial = YesNoPartial.UNKNOWN
+    docs_quality: Annotated[
+        str,
+        "Evaluation of the quality of code documentation, including API references, inline comments, and guides",
+    ] = "Unknown"
+    outdated_content: Annotated[
+        bool,
+        "Flags whether the documentation contains outdated or misleading information",
+    ] = False
+
+
+class OverallAssessment(BaseModel):
+    key_shortcomings: Annotated[
+        list[str],
+        "List of the most significant and critical issues that need to be addressed",
+    ] = ["There are no critical issues"]
+    recommendations: Annotated[
+        list[str],
+        "Specific improvements to address issues or optimize the process",
+    ] = ["No recommendations"]
+
+
+class RepositoryReport(BaseModel):
+    structure: RepositoryStructure = RepositoryStructure()
+    readme: ReadmeEvaluation = ReadmeEvaluation()
+    documentation: CodeDocumentation = CodeDocumentation()
+    assessment: OverallAssessment = OverallAssessment()
+
+    class Config:
+        extra = "ignore"