hirundo 0.1.8__py3-none-any.whl → 0.1.16__py3-none-any.whl

hirundo/dataset_optimization_results.py

@@ -0,0 +1,42 @@
+import typing
+from pathlib import Path
+
+from pydantic import BaseModel
+from typing_extensions import TypeAliasType
+
+from hirundo._dataframe import has_pandas, has_polars
+
+DataFrameType = TypeAliasType("DataFrameType", None)
+
+if has_pandas:
+    from hirundo._dataframe import pd
+
+    DataFrameType = TypeAliasType("DataFrameType", typing.Union[pd.DataFrame, None])
+if has_polars:
+    from hirundo._dataframe import pl
+
+    DataFrameType = TypeAliasType("DataFrameType", typing.Union[pl.DataFrame, None])
+
+
+T = typing.TypeVar("T")
+
+
+class DatasetOptimizationResults(BaseModel, typing.Generic[T]):
+    model_config = {"arbitrary_types_allowed": True}
+
+    cached_zip_path: Path
+    """
+    The path to the cached zip file of the results
+    """
+    suspects: T
+    """
+    A polars/pandas DataFrame containing the results of the optimization run
+    """
+    object_suspects: typing.Optional[T]
+    """
+    A polars/pandas DataFrame containing the object-level results of the optimization run
+    """
+    warnings_and_errors: T
+    """
+    A polars/pandas DataFrame containing the warnings and errors of the optimization run
+    """

hirundo/git.py

@@ -1,14 +1,15 @@
+import datetime
 import re
 import typing
-from typing import Annotated
 
 import pydantic
 import requests
 from pydantic import BaseModel, field_validator
 from pydantic_core import Url
 
+from hirundo._constraints import RepoUrl
 from hirundo._env import API_HOST
-from hirundo._headers import get_auth_headers, json_headers
+from hirundo._headers import get_headers
 from hirundo._http import raise_for_status_with_reason
 from hirundo._timeouts import MODIFY_TIMEOUT, READ_TIMEOUT
 from hirundo.logger import get_logger
@@ -16,7 +17,7 @@ from hirundo.logger import get_logger
 logger = get_logger(__name__)
 
 
-class GitPlainAuthBase(BaseModel):
+class GitPlainAuth(BaseModel):
     username: str
     """
     The username for the Git repository
@@ -27,7 +28,7 @@ class GitPlainAuthBase(BaseModel):
     """
 
 
-class GitSSHAuthBase(BaseModel):
+class GitSSHAuth(BaseModel):
     ssh_key: str
     """
     The SSH key for the Git repository
@@ -48,10 +49,10 @@ class GitRepo(BaseModel):
     """
     A name to identify the Git repository in the Hirundo system.
     """
-    repository_url: Annotated[str, Url]
+    repository_url: typing.Union[str, RepoUrl]
     """
     The URL of the Git repository, it should start with `ssh://` or `https://` or be in the form `user@host:path`.
-    If it is in the form `user@host:path`, it will be rewritten to `ssh://user@host:path`.
+    If it is in the form `user@host:path`, it will be rewritten to `ssh://user@host/path`.
     """
     organization_id: typing.Optional[int] = None
     """
@@ -59,14 +60,14 @@ class GitRepo(BaseModel):
     If not provided, it will be assigned to your default organization.
     """
 
-    plain_auth: typing.Optional[GitPlainAuthBase] = pydantic.Field(
+    plain_auth: typing.Optional[GitPlainAuth] = pydantic.Field(
         default=None, examples=[None, {"username": "ben", "password": "password"}]
     )
     """
     The plain authentication details for the Git repository.
     Use this if using a special user with a username and password for authentication.
     """
-    ssh_auth: typing.Optional[GitSSHAuthBase] = pydantic.Field(
+    ssh_auth: typing.Optional[GitSSHAuth] = pydantic.Field(
         default=None,
         examples=[
             {
@@ -84,34 +85,46 @@ class GitRepo(BaseModel):
 
     @field_validator("repository_url", mode="before", check_fields=True)
     @classmethod
-    def check_valid_repository_url(cls, repository_url: str):
-        # Check if the URL already has a protocol
-        if not re.match(r"^[a-z]+://", repository_url):
-            # Check if the URL has the `@` and `:` pattern with a non-numeric section before the next slash
-            match = re.match(r"([^@]+@[^:]+):([^0-9/][^/]*)/(.+)", repository_url)
-            if match:
-                user_host = match.group(1)
-                path = match.group(2) + "/" + match.group(3)
-                rewritten_url = f"ssh://{user_host}/{path}"
-                logger.info("Modified Git repo to add SSH protocol", rewritten_url)
-                return rewritten_url
-        if not repository_url.startswith("ssh://") and not repository_url.startswith(
-            "https://"
-        ):
+    def check_valid_repository_url(cls, repository_url: typing.Union[str, RepoUrl]):
+        # Check if the URL has the `@` and `:` pattern with a non-numeric section before the next slash
+        match = re.match("([^@]+@[^:]+):([^0-9/][^/]*)/(.+)", str(repository_url))
+        if match:
+            user_host = match.group(1)
+            path = match.group(2) + "/" + match.group(3)
+            rewritten_url = Url(f"ssh://{user_host}/{path}")
+            # Check if the URL already has a protocol
+            url_scheme = rewritten_url.scheme
+            logger.info(
+                "Modified Git repo to replace %s@%s:%s/%s with %s",
+                url_scheme,
+                match.group(1),
+                match.group(2),
+                match.group(3),
+                rewritten_url,
+            )
+            return rewritten_url
+        if not str(repository_url).startswith("ssh://") and not str(
+            repository_url
+        ).startswith("https://"):
             raise ValueError("Repository URL must start with 'ssh://' or 'https://'")
+        if not isinstance(repository_url, Url):
+            repository_url = Url(repository_url)
         return repository_url
 
-    def create(self):
+    def create(self, replace_if_exists: bool = False) -> int:
         """
         Create a Git repository in the Hirundo system.
+
+        Args:
+            replace_if_exists: If a Git repository with the same name already exists, replace it.
         """
         git_repo = requests.post(
             f"{API_HOST}/git-repo/",
-            json=self.model_dump(),
-            headers={
-                **json_headers,
-                **get_auth_headers(),
+            json={
+                **self.model_dump(mode="json"),
+                "replace_if_exists": replace_if_exists,
             },
+            headers=get_headers(),
             timeout=MODIFY_TIMEOUT,
         )
         raise_for_status_with_reason(git_repo)
@@ -120,19 +133,57 @@ class GitRepo(BaseModel):
         return git_repo_id
 
     @staticmethod
-    def list():
+    def get_by_id(git_repo_id: int) -> "GitRepoOut":
+        """
+        Retrieves a `GitRepo` instance from the server by its ID
+
+        Args:
+            git_repo_id: The ID of the `GitRepo` to retrieve
+        """
+        git_repo = requests.get(
+            f"{API_HOST}/git-repo/{git_repo_id}",
+            headers=get_headers(),
+            timeout=READ_TIMEOUT,
+        )
+        raise_for_status_with_reason(git_repo)
+        return GitRepoOut(**git_repo.json())
+
+    @staticmethod
+    def get_by_name(
+        name: str,
+    ) -> "GitRepoOut":
+        """
+        Retrieves a `GitRepo` instance from the server by its name
+
+        Args:
+            name: The name of the `GitRepo` to retrieve
+        """
+        git_repo = requests.get(
+            f"{API_HOST}/git-repo/by-name/{name}",
+            headers=get_headers(),
+            timeout=READ_TIMEOUT,
+        )
+        raise_for_status_with_reason(git_repo)
+        return GitRepoOut(**git_repo.json())
+
+    @staticmethod
+    def list() -> list["GitRepoOut"]:
         """
         List all Git repositories in the Hirundo system.
         """
         git_repos = requests.get(
             f"{API_HOST}/git-repo/",
-            headers={
-                **get_auth_headers(),
-            },
+            headers=get_headers(),
             timeout=READ_TIMEOUT,
         )
         raise_for_status_with_reason(git_repos)
-        return git_repos.json()
+        git_repo_json = git_repos.json()
+        return [
+            GitRepoOut(
+                **git_repo,
+            )
+            for git_repo in git_repo_json
+        ]
 
     @staticmethod
     def delete_by_id(git_repo_id: int):
@@ -144,9 +195,7 @@ class GitRepo(BaseModel):
         """
         git_repo = requests.delete(
             f"{API_HOST}/git-repo/{git_repo_id}",
-            headers={
-                **get_auth_headers(),
-            },
+            headers=get_headers(),
             timeout=MODIFY_TIMEOUT,
         )
         raise_for_status_with_reason(git_repo)
@@ -158,3 +207,12 @@ class GitRepo(BaseModel):
         if not self.id:
             raise ValueError("No GitRepo has been created")
         GitRepo.delete_by_id(self.id)
+
+
+class GitRepoOut(BaseModel):
+    id: int
+    name: str
+    repository_url: RepoUrl
+
+    created_at: datetime.datetime
+    updated_at: datetime.datetime