recce-nightly 1.3.0.20250507__py3-none-any.whl → 1.4.0.20250515__py3-none-any.whl

recce/git.py

@@ -1,12 +1,12 @@
 import os
 
-from git import Repo, InvalidGitRepositoryError
+from git import InvalidGitRepositoryError, Repo
 
 
 def current_default_branch():
     try:
         repo = Repo(search_parent_directories=True)
-        return repo.remotes.origin.refs['HEAD'].reference.remote_head
+        return repo.remotes.origin.refs["HEAD"].reference.remote_head
     except Exception:
         return None
 
@@ -53,19 +53,19 @@ def commit_hash_from_branch(branch: str, short_length: int = 7, short: bool = Fa
         return None
 
 
-def hosting_repo(remote: str = 'origin'):
+def hosting_repo(remote: str = "origin"):
     try:
         repo = Repo(search_parent_directories=True)
         origin_url = repo.remote(name=remote).url
         remote_repo = None
 
-        if origin_url.startswith('git@'):
+        if origin_url.startswith("git@"):
             # Handle git@github.com:user/repo.git
-            remote_repo = origin_url.split(':')[1].replace('.git', '')
+            remote_repo = origin_url.split(":")[1].replace(".git", "")
 
-        elif origin_url.startswith('https://') or origin_url.startswith('http://'):
+        elif origin_url.startswith("https://") or origin_url.startswith("http://"):
             # Handle https://github.com/user/repo.git or http://github.com/user/repo.git
-            remote_repo = '/'.join(origin_url.split('/')[-2:]).replace('.git', '')
+            remote_repo = "/".join(origin_url.split("/")[-2:]).replace(".git", "")
 
         return remote_repo
     except ValueError:

recce/github.py

@@ -3,10 +3,10 @@ import os
 import re
 import zipfile
 from datetime import datetime
-from typing import List, Tuple, Optional
+from typing import List, Optional, Tuple
 
 import requests
-from github import Artifact, Github, Auth, UnknownObjectException, PullRequest
+from github import Artifact, Auth, Github, PullRequest, UnknownObjectException
 
 from recce.git import current_branch, hosting_repo
 
@@ -16,7 +16,7 @@ def download_artifact(github_token: str, artifact: Artifact) -> List[str]:
     Download the artifact from the Github.
     """
     headers = {
-        'Authorization': f'Bearer {github_token}',
+        "Authorization": f"Bearer {github_token}",
     }
     r = requests.get(artifact.archive_download_url, headers=headers)
     r.raise_for_status()
@@ -27,17 +27,17 @@ def download_artifact(github_token: str, artifact: Artifact) -> List[str]:
 
 def recce_ci_artifact(**kwargs):
     """
-        Download the artifact from the GitHub CI.
+    Download the artifact from the GitHub CI.
     """
-    from rich.console import Console
-
     import git
+
     # Authentication is defined via github.Auth
     from github import Auth, Github
+    from rich.console import Console
 
     console = Console()
-    github_token = kwargs.get('github_token')
-    github_repository = kwargs.get('github_repository')
+    github_token = kwargs.get("github_token")
+    github_repository = kwargs.get("github_repository")
 
     if github_token is None:
         console.print("[[red]Error[/red]] Missing GitHub token. Please provide a GitHub token.")
@@ -70,11 +70,11 @@ def recce_ci_artifact(**kwargs):
 
     head_branch = github_repo.get_branch(current_branch)
 
-    console.rule('GitHub Repository Information')
-    console.print(f'Repository:     {github_repository}')
-    console.print(f'Current Branch: {current_branch}')
+    console.rule("GitHub Repository Information")
+    console.print(f"Repository:     {github_repository}")
+    console.print(f"Current Branch: {current_branch}")
 
-    pull_requests = github_repo.get_pulls(head=f'{github_owner}:{current_branch}')
+    pull_requests = github_repo.get_pulls(head=f"{github_owner}:{current_branch}")
     if pull_requests.totalCount == 0:
         # No pull request found for the current branch
         console.print("[[red]Error[/red]] No pull request found for the current branch.")
@@ -82,32 +82,32 @@ def recce_ci_artifact(**kwargs):
 
     pr = pull_requests[0]
 
-    console.rule('GitHub Pull Request Information')
-    console.print(f'{pr.title} - {pr.html_url}')
-    console.print(f'State: {pr.state.title()}')
-    console.print(f'Author: {pr.user.name}')
-    console.print(f'Created At: {pr.created_at}')
+    console.rule("GitHub Pull Request Information")
+    console.print(f"{pr.title} - {pr.html_url}")
+    console.print(f"State: {pr.state.title()}")
+    console.print(f"Author: {pr.user.name}")
+    console.print(f"Created At: {pr.created_at}")
 
     workflow_runs = github_repo.get_workflow_runs(
-        event='pull_request',
-        status='success',
+        event="pull_request",
+        status="success",
         branch=head_branch,
     )
     if workflow_runs.totalCount == 0:
         console.print("[[yellow]Skip[/yellow]] No successful workflow runs found.")
         return 0
     last_workflow_run = workflow_runs[0]
-    console.rule('GitHub Workflow Run Information')
-    console.print(f'Last Workflow Run: {last_workflow_run.name} {last_workflow_run.html_url}')
+    console.rule("GitHub Workflow Run Information")
+    console.print(f"Last Workflow Run: {last_workflow_run.name} {last_workflow_run.html_url}")
     artifacts = last_workflow_run.get_artifacts()
     if artifacts.totalCount == 0:
         console.print("[[yellow]Skip[/yellow]] No artifacts found.")
         return 0
     for artifact in artifacts:
-        console.print(f'Artifact: {artifact.name} {artifact.archive_download_url}')
+        console.print(f"Artifact: {artifact.name} {artifact.archive_download_url}")
         artifact_files = download_artifact(github_token, artifact)
         console.print(f'Extracted Files: {", ".join(artifact_files)}')
-    console.rule('Complete')
+    console.rule("Complete")
     return 0
 
 
@@ -121,7 +121,7 @@ def get_pull_request(branch, owner, repo_name, github_token=None) -> Tuple[Optio
 
     try:
         repo = g.get_repo(f"{owner}/{repo_name}")
-        pulls = repo.get_pulls(state='open')
+        pulls = repo.get_pulls(state="open")
 
         for pr in pulls:
             if pr.head.ref == branch:
@@ -129,7 +129,10 @@ def get_pull_request(branch, owner, repo_name, github_token=None) -> Tuple[Optio
 
     except UnknownObjectException:
         if github_token is not None:
-            return None, f"Repository {owner}/{repo_name} not found. If it is private repo, please add the 'repo' scope to the token."
+            return (
+                None,
+                f"Repository {owner}/{repo_name} not found. If it is private repo, please add the 'repo' scope to the token.",
+            )
 
     return None, None
 
@@ -139,53 +142,54 @@ def recce_pr_information(github_token=None) -> Tuple[Optional[type(PullRequest)]
     repo = hosting_repo()
 
     if not repo:
-        return None, 'This is not a git repository.'
-    if '/' not in repo:
-        return None, 'This is not a GitHub repository.'
+        return None, "This is not a git repository."
+    if "/" not in repo:
+        return None, "This is not a GitHub repository."
 
-    owner, repo_name = repo.split('/')
+    owner, repo_name = repo.split("/")
 
     github_token = github_token if github_token else os.getenv("GITHUB_TOKEN")
     return get_pull_request(branch, owner, repo_name, github_token)
 
 
 def is_github_codespace():
-    return os.getenv('CODESPACES') == 'true'
+    return os.getenv("CODESPACES") == "true"
 
 
 def get_github_codespace_name():
-    return os.getenv('CODESPACE_NAME')
+    return os.getenv("CODESPACE_NAME")
 
 
 def get_github_codespace_info():
     if is_github_codespace() is False:
         return None
 
-    codespace_name = os.environ.get('CODESPACE_NAME')
-    github_token = os.environ.get('GITHUB_TOKEN')
+    codespace_name = os.environ.get("CODESPACE_NAME")
+    github_token = os.environ.get("GITHUB_TOKEN")
 
     response = requests.get(
-        f'https://api.github.com/user/codespaces/{codespace_name}',
+        f"https://api.github.com/user/codespaces/{codespace_name}",
         headers={
-            'Accept': 'application/vnd.github+json',
-            'Authorization': f'token {github_token}',
-            'X-GitHub-Api-Version': '2022-11-28',
-        })
+            "Accept": "application/vnd.github+json",
+            "Authorization": f"token {github_token}",
+            "X-GitHub-Api-Version": "2022-11-28",
+        },
+    )
 
     if response.status_code != 200:
         return None
     codespace_info = response.json()
 
     return dict(
-        name=codespace_info.get('name'),
-        machine=codespace_info.get('machine'),
-        prebuild=codespace_info.get('prebuild'),
-        created_at=codespace_info.get('created_at'),
-        updated_at=codespace_info.get('updated_at'),
-        last_used_at=codespace_info.get('last_used_at'),
-        state=codespace_info.get('state'),
-        location=codespace_info.get('location'),
-        idle_timeout_minutes=codespace_info.get('idle_timeout_minutes'),
+        name=codespace_info.get("name"),
+        machine=codespace_info.get("machine"),
+        prebuild=codespace_info.get("prebuild"),
+        created_at=codespace_info.get("created_at"),
+        updated_at=codespace_info.get("updated_at"),
+        last_used_at=codespace_info.get("last_used_at"),
+        state=codespace_info.get("state"),
+        location=codespace_info.get("location"),
+        idle_timeout_minutes=codespace_info.get("idle_timeout_minutes"),
     )
 
 
@@ -194,25 +198,25 @@ def get_github_codespace_available_at(codespace):
         return None
 
     def search_in_file(file_path, search_string):
-        with open(file_path, 'r') as f:
+        with open(file_path, "r") as f:
             for _, line in enumerate(f, 1):
                 if search_string in line:
                     return line
         return None
 
     def extract_datatime(log_line):
-        pattern = r'\[(\d{4}-\d{2}-\d{2} \d{2}:\d{2}:\d{2}\.\d{3}).*\]'
+        pattern = r"\[(\d{4}-\d{2}-\d{2} \d{2}:\d{2}:\d{2}\.\d{3}).*\]"
         match = re.search(pattern, log_line)
         if match:
             datetime_str = match.group(1)
-            return datetime.strptime(datetime_str, '%Y-%m-%d %H:%M:%S.%f')
+            return datetime.strptime(datetime_str, "%Y-%m-%d %H:%M:%S.%f")
         return None
 
-    github_codepsce_log_dir = '/tmp/codespaces_logs'
+    github_codepsce_log_dir = "/tmp/codespaces_logs"
     try:
         log_file = os.listdir(github_codepsce_log_dir)[-1]  # Get the latest log file
-        start_monitor_line = search_in_file(f'{github_codepsce_log_dir}/{log_file}', 'Starting monitor')
+        start_monitor_line = search_in_file(f"{github_codepsce_log_dir}/{log_file}", "Starting monitor")
         return extract_datatime(start_monitor_line)
     except Exception:
         # If there is any error, use the updated_at time from the codespace info
-        return datetime.fromisoformat(codespace.get('updated_at'))
+        return datetime.fromisoformat(codespace.get("updated_at"))

recce/models/__init__.py

@@ -1,3 +1,3 @@
 from .check import CheckDAO
 from .run import RunDAO
-from .types import Check, Run, RunType, RunProgress
+from .types import Check, Run, RunProgress, RunType

recce/models/check.py

@@ -1,6 +1,7 @@
-from typing import Optional, List
+from typing import List, Optional
 
 from recce.exceptions import RecceException
+
 from .types import Check
 
 
@@ -12,6 +13,7 @@ class CheckDAO:
     @property
     def _checks(self):
         from recce.core import default_context
+
         return default_context().checks
 
     def create(self, check) -> None:
@@ -38,10 +40,10 @@ class CheckDAO:
     def reorder(self, source: int, destination: int):
 
         if source < 0 or source >= len(self._checks):
-            raise RecceException('Failed to reorder checks. Source index out of range')
+            raise RecceException("Failed to reorder checks. Source index out of range")
 
         if destination < 0 or destination >= len(self._checks):
-            raise RecceException('Failed to reorder checks. Destination index out of range')
+            raise RecceException("Failed to reorder checks. Destination index out of range")
 
         check_to_move = self._checks.pop(source)
         self._checks.insert(destination, check_to_move)
@@ -50,7 +52,4 @@ class CheckDAO:
         self._checks.clear()
 
     def status(self):
-        return {
-            'total': len(self._checks),
-            'approved': len([c for c in self._checks if c.is_checked])
-        }
+        return {"total": len(self._checks), "approved": len([c for c in self._checks if c.is_checked])}

recce/models/run.py

@@ -9,6 +9,7 @@ class RunDAO:
     @property
     def _runs(self):
         from recce.core import default_context
+
         return default_context().runs
 
     def create(self, run: Run):

recce/models/types.py

@@ -1,26 +1,26 @@
 import uuid
 from datetime import datetime, timezone
 from enum import Enum
-from typing import Optional, Literal
+from typing import Literal, Optional
 
-from pydantic import BaseModel, UUID4, Field
+from pydantic import UUID4, BaseModel, Field
 
 
 class RunType(Enum):
-    SIMPLE = 'simple'
+    SIMPLE = "simple"
     QUERY = "query"
-    QUERY_BASE = 'query_base'
-    QUERY_DIFF = 'query_diff'
-    VALUE_DIFF = 'value_diff'
-    VALUE_DIFF_DETAIL = 'value_diff_detail'
-    SCHEMA_DIFF = 'schema_diff'
-    PROFILE = 'profile'
-    PROFILE_DIFF = 'profile_diff'
-    ROW_COUNT = 'row_count'
-    ROW_COUNT_DIFF = 'row_count_diff'
-    LINEAGE_DIFF = 'lineage_diff'
-    TOP_K_DIFF = 'top_k_diff'
-    HISTOGRAM_DIFF = 'histogram_diff'
+    QUERY_BASE = "query_base"
+    QUERY_DIFF = "query_diff"
+    VALUE_DIFF = "value_diff"
+    VALUE_DIFF_DETAIL = "value_diff_detail"
+    SCHEMA_DIFF = "schema_diff"
+    PROFILE = "profile"
+    PROFILE_DIFF = "profile_diff"
+    ROW_COUNT = "row_count"
+    ROW_COUNT_DIFF = "row_count_diff"
+    LINEAGE_DIFF = "lineage_diff"
+    TOP_K_DIFF = "top_k_diff"
+    HISTOGRAM_DIFF = "histogram_diff"
 
     def __str__(self):
         return self.value
@@ -32,12 +32,12 @@ class RunProgress(BaseModel):
 
 
 class RunStatus(Enum):
-    FINISHED = 'finished'
-    FAILED = 'failed'
-    CANCELLED = 'cancelled'
-    RUNNING = 'running'
+    FINISHED = "finished"
+    FAILED = "failed"
+    CANCELLED = "cancelled"
+    RUNNING = "running"
     # This is a special status only in v0.36.0. Replaced by FINISHED. To be removed in the future.
-    SUCCESSFUL = 'successful'
+    SUCCESSFUL = "successful"
 
 
 class Run(BaseModel):
@@ -84,15 +84,15 @@ class Check(BaseModel):
 
 
 ChangeStatus = Literal[
-    'added',
-    'removed',
-    'modified',
+    "added",
+    "removed",
+    "modified",
 ]
 ChangeCategory = Literal[
-    'breaking',
-    'non_breaking',
-    'partial_breaking',
-    'unknown',
+    "breaking",
+    "non_breaking",
+    "partial_breaking",
+    "unknown",
 ]
 
 

recce/pull_request.py

@@ -28,14 +28,14 @@ def fetch_pr_metadata(**kwargs):
     # fetch from github action event path
     metadata = fetch_pr_metadata_from_event_path()
     if metadata is not None:
-        pr_info.id = metadata.get('github_pr_id')
-        pr_info.url = metadata.get('github_pr_url')
-        pr_info.title = metadata.get('github_pr_title')
-        pr_info.repository = metadata.get('github_repository')
+        pr_info.id = metadata.get("github_pr_id")
+        pr_info.url = metadata.get("github_pr_url")
+        pr_info.title = metadata.get("github_pr_title")
+        pr_info.repository = metadata.get("github_repository")
     else:
         repo = hosting_repo()
-        pr, message = recce_pr_information(github_token=kwargs.get('github_token'))
-        if kwargs.get('cloud') and message:
+        pr, message = recce_pr_information(github_token=kwargs.get("github_token"))
+        if kwargs.get("cloud") and message:
             print(message)
 
         if pr:
@@ -47,14 +47,14 @@ def fetch_pr_metadata(**kwargs):
             pr_info.branch = pr.head.ref
 
     # fetch from cli arguments
-    if pr_info.url is None and 'github_pull_request_url' in kwargs:
-        pr_info.url = kwargs.get('github_pull_request_url')
+    if pr_info.url is None and "github_pull_request_url" in kwargs:
+        pr_info.url = kwargs.get("github_pull_request_url")
 
     if pr_info.branch is None:
-        pr_info.branch = kwargs.get('git_current_branch')
+        pr_info.branch = kwargs.get("git_current_branch")
 
     if pr_info.base_branch is None:
-        pr_info.base_branch = kwargs.get('git_base_branch')
+        pr_info.base_branch = kwargs.get("git_base_branch")
 
     # fetch from env
     if pr_info.url is None:
@@ -65,17 +65,17 @@ def fetch_pr_metadata(**kwargs):
 
 def fetch_pr_metadata_from_event_path() -> Optional[dict]:
     """
-        If recce run is running in a GitHub Action, this function will return the pull request metadata.
+    If recce run is running in a GitHub Action, this function will return the pull request metadata.
 
-        Example:
-        {
-            "github_pr_id": 1,
-            "github_pr_url": "https://github.com/xyz/abc/pull/1,
-            "github_pr_title": "Update README.md",
-            "github_repository": "xyz/abc"
-        }
+    Example:
+    {
+        "github_pr_id": 1,
+        "github_pr_url": "https://github.com/xyz/abc/pull/1,
+        "github_pr_title": "Update README.md",
+        "github_repository": "xyz/abc"
+    }
 
-        :return: dict
+    :return: dict
     """
 
     # get the event json from the path in GITHUB_EVENT_PATH
@@ -92,10 +92,12 @@ def fetch_pr_metadata_from_event_path() -> Optional[dict]:
                 pr_url = pull_request_data["_links"]["html"]["href"]
                 pr_api = pull_request_data["_links"]["self"]["href"]
                 pr_title = _fetch_pr_title(pr_api)
-                return dict(github_pr_id=pr_id,
-                            github_pr_url=pr_url,
-                            github_pr_title=pr_title,
-                            github_repository=github_repository)
+                return dict(
+                    github_pr_id=pr_id,
+                    github_pr_url=pr_url,
+                    github_pr_title=pr_title,
+                    github_repository=github_repository,
+                )
             else:
                 print("Not a pull request event, skip.")
         except Exception as e:
@@ -115,7 +117,7 @@ def _fetch_pr_title(endpoint) -> Optional[str]:
 
         if response.status_code == 200:
             pull_request_data = response.json()
-            return pull_request_data.get('title')
+            return pull_request_data.get("title")
     except Exception as e:
         print("Cannot fetch PR title: ", e)