feat(routes): add retrieve file content route

README.md

@@ -6,8 +6,6 @@ Pytest plugin for automatically mocking OpenAI requests. Powered by [RESPX](http
 
 ## Supported Endpoints
 
-[![support coverage](https://img.shields.io/badge/Endpoints_Covered-49.2%25-orange)](https://mharrisb1.github.io/openai-responses-python/coverage)
-
 - [Chat](https://platform.openai.com/docs/api-reference/chat)
 - [Embeddings](https://platform.openai.com/docs/api-reference/embeddings)
 - [Files](https://platform.openai.com/docs/api-reference/files)

docs/coverage.md

@@ -37,7 +37,7 @@ The end-goal of this library is to eventually support all OpenAI API routes. See
 | List files                         | :material-check:{ .green } |             -              | Stateful   |
 | Retrieve file                      | :material-check:{ .green } |             -              | Stateful   |
 | Delete file                        | :material-check:{ .green } |             -              | Stateful   |
-| Retrieve file content              |  :material-close:{ .red }  |             -              | Stateful   |
+| Retrieve file content              | :material-check:{ .green } |             -              | Stateful   |
 | **Images**                         |
 | Create image                       |  :material-close:{ .red }  |             -              | -          |
 | Create image edit                  |  :material-close:{ .red }  |             -              | -          |

examples/test_files.py

@@ -64,3 +64,15 @@ def test_delete_file(openai_mock: OpenAIMock):
     assert client.files.delete(file.id).deleted
     assert openai_mock.files.create.route.call_count == 1
     assert openai_mock.files.delete.route.call_count == 1
+
+
+@openai_responses.mock()
+def test_retrieve_file_content(openai_mock: OpenAIMock):
+    client = openai.Client(api_key="sk-fake123")
+
+    file = client.files.create(
+        file=open("examples/example.json", "rb"),
+        purpose="fine-tune",
+    )
+    res = client.files.content(file.id)
+    assert res.content == b'{\n  "foo": "bar",\n  "fizz": "buzz"\n}'

pyproject.toml

@@ -23,6 +23,7 @@ python = ">=3.9,<4.0"
 openai = "^1.25"
 respx = "^0.20.2"
 faker-openai-api-provider = "^0.1.0"
+requests-toolbelt = "^1.0.0"
 
 [tool.poetry.group.dev.dependencies]
 black = "^24.2.0"
@@ -37,6 +38,9 @@ tox = "^4.14.2"
 requires = ["poetry-core"]
 build-backend = "poetry.core.masonry.api"
 
+[tool.mypy]
+ignore_missing_imports = true
+
 [tool.pyright]
 typeCheckingMode = "strict"
 pythonVersion = "3.9"

src/openai_responses/_routes/__init__.py

@@ -4,7 +4,13 @@
 
 from .chat import ChatCompletionsCreateRoute
 from .embeddings import EmbeddingsCreateRoute
-from .files import FileCreateRoute, FileListRoute, FileRetrieveRoute, FileDeleteRoute
+from .files import (
+    FileCreateRoute,
+    FileListRoute,
+    FileRetrieveRoute,
+    FileDeleteRoute,
+    FileRetrieveContentRoute,
+)
 from .models import ModelListRoute, ModelRetrieveRoute
 from .assistants import (
     AssistantCreateRoute,
@@ -66,6 +72,7 @@ def __init__(self, router: respx.MockRouter, state: StateStore) -> None:
         self.list = FileListRoute(router, state)
         self.retrieve = FileRetrieveRoute(router, state)
         self.delete = FileDeleteRoute(router, state)
+        self.content = FileRetrieveContentRoute(router, state)
 
 
 class ModelRoutes:

src/openai_responses/_routes/files.py

@@ -1,9 +1,10 @@
-import re
-from typing import Any
+import sys
+from typing import Any, Optional
 from typing_extensions import override
 
 import httpx
 import respx
+from requests_toolbelt.multipart import decoder
 
 from openai.pagination import SyncPage
 from openai.types.file_object import FileObject
@@ -22,7 +23,6 @@
 from .._utils.serde import model_dict
 from .._utils.time import utcnow_unix_timestamp_s
 
-REGEXP_FILE = r'Content-Disposition: form-data;[^;]+; name="purpose"\r\n\r\n(?P<purpose_value>[^\r\n]+)|filename="(?P<filename>[^"]+)"'
 
 __all__ = ["FileCreateRoute", "FileListRoute", "FileRetrieveRoute", "FileDeleteRoute"]
 
@@ -38,39 +38,51 @@ def __init__(self, router: respx.MockRouter, state: StateStore) -> None:
     @override
     def _handler(self, request: httpx.Request, route: respx.Route) -> httpx.Response:
         self._route = route
-        model = self._build({}, request)
+
+        filename: Optional[str] = None
+        purpose: Optional[str] = None
+        file_content: Optional[bytes] = None
+
+        multipart_data = decoder.MultipartDecoder(
+            request.content, request.headers.get("content-type")
+        )
+        for part in multipart_data.parts:  # type: ignore
+            content_disposition = part.headers.get(b"Content-Disposition", b"").decode()  # type: ignore
+            if 'name="purpose"' in content_disposition:
+                purpose = part.text  # type: ignore
+            elif 'name="file"' in content_disposition:
+                filename = (  # type: ignore
+                    part.headers.get(b"Content-Disposition", b"")  # type: ignore
+                    .decode()
+                    .split("filename=")[1]
+                    .strip('"')
+                )
+                file_content = part.content  # type: ignore
+
+        assert filename
+        assert purpose
+        assert file_content
+
+        model = FileObject(
+            id=faker.file.id(),
+            bytes=sys.getsizeof(file_content),  # type: ignore
+            created_at=utcnow_unix_timestamp_s(),
+            filename=filename,  # type: ignore
+            object="file",
+            purpose=purpose,  # type: ignore
+            status="uploaded",
+            status_details=None,
+        )
         self._state.files.put(model)
+        self._state.files.content.put(model.id, file_content)  # type: ignore
         return httpx.Response(
             status_code=self._status_code,
             json=model_dict(model),
         )
 
     @staticmethod
     def _build(partial: PartialFileObject, request: httpx.Request) -> FileObject:
-        content = request.content.decode("utf-8")
-
-        filename = ""
-        purpose = "assistants"
-
-        # FIXME: hacky
-        prog = re.compile(REGEXP_FILE)
-        matches = prog.finditer(content)
-        for match in matches:
-            if match.group("filename"):
-                filename = match.group("filename")
-            if match.group("purpose_value"):
-                purpose = match.group("purpose_value")
-
-        return FileObject(
-            id=partial.get("id", faker.file.id()),
-            bytes=partial.get("bytes", 0),
-            created_at=partial.get("created_at", utcnow_unix_timestamp_s()),
-            filename=partial.get("filename", filename),
-            object="file",
-            purpose=partial.get("purpose", purpose),  # type: ignore
-            status=partial.get("status", "uploaded"),
-            status_details=partial.get("status_details"),
-        )
+        raise NotImplementedError
 
 
 class FileListRoute(StatefulRoute[SyncPage[FileObject], PartialFileList]):
@@ -159,3 +171,33 @@ def _build(
         request: httpx.Request,
     ) -> FileObject:
         raise NotImplementedError
+
+
+class FileRetrieveContentRoute(StatefulRoute[FileObject, PartialFileObject]):
+    def __init__(self, router: respx.MockRouter, state: StateStore) -> None:
+        super().__init__(
+            route=router.get(url__regex=r"/files/(?P<file_id>[a-zA-Z0-9\-]+)/content"),
+            status_code=200,
+            state=state,
+        )
+
+    @override
+    def _handler(
+        self,
+        request: httpx.Request,
+        route: respx.Route,
+        **kwargs: Any,
+    ) -> httpx.Response:
+        self._route = route
+        fil_id = kwargs["file_id"]
+        found = self._state.files.get(fil_id)
+        if not found:
+            return httpx.Response(404)
+
+        content = self._state.files.content.get(found.id)
+        assert content
+        return httpx.Response(status_code=200, content=content)
+
+    @staticmethod
+    def _build(partial: PartialFileObject, request: httpx.Request) -> FileObject:
+        raise NotImplementedError

src/openai_responses/stores/__init__.py

@@ -1,3 +1,4 @@
+from .content_store import ContentStore
 from .state_store import StateStore
 
-__all__ = ["StateStore"]
+__all__ = ["StateStore", "ContentStore"]

src/openai_responses/stores/content_store.py

@@ -0,0 +1,15 @@
+from typing import Dict, Union
+
+
+class ContentStore:
+    def __init__(self) -> None:
+        self._data: Dict[str, bytes] = {}
+
+    def put(self, id: str, content: bytes) -> None:
+        self._data[id] = content
+
+    def get(self, id: str) -> Union[bytes, None]:
+        return self._data.get(id)
+
+    def delete(self, id: str) -> None:
+        del self._data[id]

src/openai_responses/stores/state_store.py

@@ -7,6 +7,7 @@
 from openai.types.beta.threads.run import Run
 from openai.types.beta.threads.runs.run_step import RunStep
 
+from .content_store import ContentStore
 from .._constants import SYSTEM_MODELS
 from .._utils.serde import model_parse
 
@@ -77,8 +78,11 @@ def delete(self, id: str) -> bool:
             return False
 
 
-# TODO: add content storage
 class FileStore(BaseStore[FileObject]):
+    def __init__(self) -> None:
+        super().__init__()
+        self.content = ContentStore()
+
     def list(
         self,
         purpose: Optional[