Vendor Python packages (take 2)

This PR vendors all Python packages using the [vendoring](https://pypi.org/project/vendoring/) library. Vendoring allows users to install packages that rely on versions of libraries that are in conflict with Cog's dependencies (e.g. Pydantic, see replicate#1562, replicate#1384, replicate#1186, replicate#1586, replicate#1336, replicate#785). Vendored packages are gitignored and are sync'd by the default `make` rule. Closes replicate#409 Signed-off-by: andreasjansson <[email protected]>
andreasjansson · Apr 16, 2024 · f6deb3a · f6deb3a
1 parent 72398c7
commit f6deb3a
Show file tree

Hide file tree

Showing 36 changed files with 425 additions and 81 deletions.
diff --git a/.gitignore b/.gitignore
@@ -15,3 +15,4 @@ pkg/dockerfile/embed/cog.whl
 .DS_Store
 docs/README.md
 docs/CONTRIBUTING.md
+python/cog/_vendor/
diff --git a/CONTRIBUTING.md b/CONTRIBUTING.md
@@ -207,6 +207,23 @@ To publish a prerelease version, append a [SemVer prerelease identifer](https://
     git tag -a v0.1.0-alpha -m "Prerelease v0.1.0"
     git push --tags
 
+## Vendoring Python packages
+
+We vendor the few Python libraries we depend on to avoid dependency hell. We use [vendoring](https://pypi.org/project/vendoring/), the same tool used by pip.
+
+The vendored packages are defined in `python/cog/_vendor/vendor.txt`. If you add/change anything in there, the default `make` rule will sync all vendored packages.
+
+If you run into issues you may need to manually add a patch. Patches are stored in `python/tools/vendoring/patches`.
+
+To create a new patch, copy the file you want to patch to `$FILENAME.new`, and edit that file as needed. For example:
+
+    cp python/cog/_vendor/curio/__main__.py python/cog/_vendor/curio/__main__.py.new
+
+Then `diff` the new and old version to a patch file in `python/tools/vendoring/patches/$PACKAGE.patch` and re-run vendoring. For example:
+
+    diff -u ./python/cog/_vendor/curio/__main__.py ./python/cog/_vendor/curio/__main__.py.new >> python/tools/vendoring/patches/curio.patch
+    make
+
 ## Troubleshooting
 
 ### `cog command not found`

diff --git a/Makefile b/Makefile
@@ -15,16 +15,21 @@ PYTHON ?= python
 PYTEST := $(PYTHON) -m pytest
 PYRIGHT := $(PYTHON) -m pyright
 RUFF := $(PYTHON) -m ruff
+PIP := $(PYTHON) -m pip
+VENDORING := $(PYTHON) -m vendoring
 
 default: all
 
 .PHONY: all
 all: cog
 
-pkg/dockerfile/embed/cog.whl: python/* python/cog/* python/cog/server/* python/cog/command/*
+python/cog/_vendor: python/cog/vendor.txt
+	$(VENDORING) sync
+
+pkg/dockerfile/embed/cog.whl: python/* python/cog/* python/cog/server/* python/cog/command/* python/cog/_vendor
 	@echo "Building Python library"
 	rm -rf dist
-	$(PYTHON) -m pip install build && $(PYTHON) -m build --wheel
+	$(PIP) install build && $(PYTHON) -m build --wheel
 	mkdir -p pkg/dockerfile/embed
 	cp dist/*.whl $@
 
@@ -68,7 +73,7 @@ test-integration: cog
 	cd test-integration/ && $(MAKE) PATH="$(PWD):$(PATH)" test
 
 .PHONY: test-python
-test-python:
+test-python: python/cog/_vendor
 	$(PYTEST) -n auto -vv --cov=python/cog  --cov-report term-missing  python/tests $(if $(FILTER),-k "$(FILTER)",)
 
 .PHONY: test
@@ -99,7 +104,7 @@ lint-go:
 	$(GO) run github.com/golangci/golangci-lint/cmd/golangci-lint run ./...
 
 .PHONY: lint-python
-lint-python:
+lint-python: python/cog/_vendor
 	$(RUFF) python/cog
 	$(RUFF) format --check python
 	@$(PYTHON) -c 'import sys; sys.exit("Warning: python >=3.10 is needed (not installed) to pass linting (pyright)") if sys.version_info < (3, 10) else None'
@@ -114,12 +119,12 @@ mod-tidy:
 
 .PHONY: install-python # install dev dependencies
 install-python:
-	$(PYTHON) -m pip install '.[dev]'
-
+	$(PIP) install '.[dev]'
+	$(PIP) install --no-deps --ignore-requires-python vendoring==1.2.0
 
 .PHONY: run-docs-server
 run-docs-server:
-	pip install mkdocs-material
+	$(PIP) install mkdocs-material
 	sed 's/docs\///g' README.md > ./docs/README.md
 	cp CONTRIBUTING.md ./docs/
 	mkdocs serve
diff --git a/pyproject.toml b/pyproject.toml
@@ -12,22 +12,12 @@ urls."Source" = "https://github.com/replicate/cog"
 
 requires-python = ">=3.7"
 dependencies = [
-  # intentionally loose. perhaps these should be vendored to not collide with user code?
-  "attrs>=20.1,<24",
-  "fastapi>=0.75.2,<0.99.0",
-  "pydantic>=1.9,<2",
-  "PyYAML",
-  "requests>=2,<3",
-  "structlog>=20,<25",
   'typing-compat; python_version < "3.8"',
-  "typing_extensions>=4.4.0",
-  "uvicorn[standard]>=0.12,<1",
 ]
 
 optional-dependencies = { "dev" = [
   "black",
   "build",
-  "httpx",
   'hypothesis<6.80.0; python_version < "3.8"',
   'hypothesis; python_version >= "3.8"',
   'numpy<1.22.0; python_version < "3.8"',
@@ -41,6 +31,17 @@ optional-dependencies = { "dev" = [
   "pytest-cov",
   "responses",
   "ruff",
+
+  # vendoring's dependencies are listed below. vendoring is pinned to
+  # python 3.8 and you can't specify --no-deps or
+  # --ignore-requires-python in pyproject.toml, so unfortunately we
+  # have to install vendoring in the Makefile.
+  "click",
+  "jsonschema",
+  "packaging",
+  "requests",
+  "rich",
+  "toml",
 ] }
 
 dynamic = ["version"]
@@ -54,7 +55,7 @@ strictParameterNoneValue = false
 # legacy behavior, fixed in PEP688
 disableBytesTypePromotions = true
 include = ["python"]
-exclude = ["python/tests"]
+exclude = ["python/tests", "python/cog/_vendor", "python/build"]
 reportMissingParameterType = "error"
 reportUnknownLambdaType = "error"
 reportUnnecessaryIsInstance = "warning"
@@ -92,6 +93,7 @@ ignore = [
 extend-exclude = [
   "python/tests/server/fixtures/*",
   "test-integration/test_integration/fixtures/*",
+  "python/cog/_vendor/*",
 ]
 
 [tool.ruff.per-file-ignores]
@@ -111,3 +113,39 @@ extend-exclude = [
   "S607", # Starting a process with a partial executable path"
   "ANN",
 ]
+
+[tool.black]
+exclude = '(\.eggs|\.git|\.hg|\.mypy|_cache|\.nox|\.tox|\.venv|\.svn|_build|buck-out|build|dist|_vendor)'
+
+[tool.vendoring]
+destination = "python/cog/_vendor/"
+requirements = "python/cog/vendor.txt"
+namespace = "cog._vendor"
+
+protected-files = ["__init__.py", "README.rst", "vendor.txt"]
+patches-dir = "python/tools/vendoring/patches"
+
+[tool.vendoring.transformations]
+substitute = []
+drop = [
+  # we won't run any dependencies as scripts
+  "bin/",
+
+  # pyyaml falls back to plain python
+  "_yaml.*.so",
+
+  # sniffio ships with tests
+  "sniffio/_tests",
+
+  # h11 ships with tests
+  "h11/tests",
+
+  # pydantic hypothesis plugin, fall back to python
+  "pydantic/_hypothesis_plugin.*.so",
+]
+
+[tool.vendoring.typing-stubs]
+
+[tool.vendoring.license.directories]
+
+[tool.vendoring.license.fallback-urls]
diff --git a/python/cog/__init__.py b/python/cog/__init__.py
@@ -1,5 +1,4 @@
-from pydantic import BaseModel
-
+from ._vendor.pydantic import BaseModel
 from .predictor import BasePredictor
 from .types import ConcatenateIterator, File, Input, Path, Secret
 

diff --git a/python/cog/files.py b/python/cog/files.py
@@ -4,7 +4,7 @@
 import os
 from urllib.parse import urlparse
 
-import requests
+from ._vendor import requests
 
 
 def upload_file(fh: io.IOBase, output_file_prefix: str = None) -> str:

diff --git a/python/cog/json.py b/python/cog/json.py
@@ -4,8 +4,7 @@
 from types import GeneratorType
 from typing import Any, Callable
 
-from pydantic import BaseModel
-
+from ._vendor.pydantic import BaseModel
 from .types import Path
 
 

diff --git a/python/cog/logging.py b/python/cog/logging.py
@@ -1,8 +1,8 @@
 import logging
 import os
 
-import structlog
-from structlog.typing import EventDict
+from ._vendor import structlog
+from ._vendor.structlog.typing import EventDict
 
 
 def replace_level_with_severity(

diff --git a/python/cog/predictor.py b/python/cog/predictor.py
@@ -21,22 +21,21 @@
 )
 from unittest.mock import patch
 
-import structlog
-
 import cog.code_xforms as code_xforms
 
+from ._vendor import structlog
+from ._vendor.typing_extensions import Annotated
+
 try:
     from typing import get_args, get_origin
 except ImportError:  # Python < 3.8
     from typing_compat import get_args, get_origin  # type: ignore
 
-import yaml
-from pydantic import BaseModel, Field, create_model
-from pydantic.fields import FieldInfo
+from ._vendor import yaml
+from ._vendor.pydantic import BaseModel, Field, create_model
+from ._vendor.pydantic.fields import FieldInfo
 
 # Added in Python 3.9. Can be from typing if we drop support for <3.9
-from typing_extensions import Annotated
-
 from .errors import ConfigDoesNotExist, PredictorNotSet
 from .types import (
     File as CogFile,
@@ -158,7 +157,7 @@ def load_config() -> Dict[str, Any]:
     config_path = os.path.abspath("cog.yaml")
     try:
         with open(config_path) as fh:
-            config = yaml.safe_load(fh)
+            config: Dict[str, Any] = yaml.safe_load(fh)  # type: ignore
     except FileNotFoundError as e:
         raise ConfigDoesNotExist(
             f"Could not find {config_path}",

diff --git a/python/cog/schema.py b/python/cog/schema.py
@@ -7,7 +7,7 @@
 from enum import Enum
 from types import ModuleType
 
-import pydantic
+from ._vendor import pydantic
 
 BUNDLED_SCHEMA_PATH = ".cog/schema.py"
 

diff --git a/python/cog/server/eventtypes.py b/python/cog/server/eventtypes.py
@@ -1,6 +1,6 @@
 from typing import Any, Dict
 
-from attrs import define, field, validators
+from .._vendor.attrs import define, field, validators
 
 
 # From worker parent process

diff --git a/python/cog/server/http.py b/python/cog/server/http.py
@@ -25,17 +25,14 @@
 if TYPE_CHECKING:
     from typing import ParamSpec
 
-import attrs
-import structlog
-import uvicorn
-from fastapi import Body, FastAPI, Header, HTTPException, Path, Response
-from fastapi.encoders import jsonable_encoder
-from fastapi.exceptions import RequestValidationError
-from fastapi.responses import JSONResponse
-from pydantic import ValidationError
-from pydantic.error_wrappers import ErrorWrapper
-
 from .. import schema
+from .._vendor import attrs, structlog, uvicorn
+from .._vendor.fastapi import Body, FastAPI, Header, HTTPException, Path, Response
+from .._vendor.fastapi.encoders import jsonable_encoder
+from .._vendor.fastapi.exceptions import RequestValidationError
+from .._vendor.fastapi.responses import JSONResponse
+from .._vendor.pydantic import ValidationError
+from .._vendor.pydantic.error_wrappers import ErrorWrapper
 from ..errors import PredictorNotSet
 from ..files import upload_file
 from ..json import upload_files
@@ -296,7 +293,7 @@ async def predict_idempotent(
                             "prediction ID must match the ID supplied in the URL"
                         ),
                         ("body", "id"),
-                    )
+                    )  # type: ignore
                 ]
             )
 
@@ -310,7 +307,7 @@ async def predict_idempotent(
         return _predict(request=request, respond_async=respond_async)
 
     def _predict(
-        *, request: PredictionRequest, respond_async: bool = False
+        *, request: Optional[PredictionRequest], respond_async: bool = False
     ) -> Response:
         # [compat] If no body is supplied, assume that this model can be run
         # with empty input. This will throw a ValidationError if that's not

diff --git a/python/cog/server/runner.py b/python/cog/server/runner.py
@@ -7,14 +7,12 @@
 from multiprocessing.pool import AsyncResult, ThreadPool
 from typing import Any, Callable, Optional, Tuple, Union, cast
 
-import requests
-import structlog
-from attrs import define
-from fastapi.encoders import jsonable_encoder
-from requests.adapters import HTTPAdapter
-from requests.packages.urllib3.util.retry import Retry  # type: ignore
-
 from .. import schema, types
+from .._vendor import requests, structlog
+from .._vendor.attrs import define
+from .._vendor.fastapi.encoders import jsonable_encoder
+from .._vendor.requests.adapters import HTTPAdapter
+from .._vendor.urllib3.util.retry import Retry  # type: ignore
 from ..files import put_file_to_signed_endpoint
 from ..json import upload_files
 from .eventtypes import Done, Heartbeat, Log, PredictionOutput, PredictionOutputType
@@ -454,7 +452,7 @@ def _make_file_upload_http_client() -> requests.Session:
             backoff_factor=0.1,
             status_forcelist=[408, 429, 500, 502, 503, 504],
             allowed_methods=["PUT"],
-        ),
+        ),  # type: ignore
     )
     session.mount("http://", adapter)
     session.mount("https://", adapter)

diff --git a/python/cog/server/webhook.py b/python/cog/server/webhook.py
@@ -1,11 +1,9 @@
 import os
 from typing import Any, Callable, Set
 
-import requests
-import structlog
-from requests.adapters import HTTPAdapter
-from requests.packages.urllib3.util.retry import Retry  # type: ignore
-
+from .._vendor import requests, structlog
+from .._vendor.requests.adapters import HTTPAdapter
+from .._vendor.urllib3.util.retry import Retry  # type: ignore
 from ..schema import Status, WebhookEvent
 from .response_throttler import ResponseThrottler
 
@@ -96,7 +94,7 @@ def requests_session_with_retries() -> requests.Session:
             backoff_factor=0.1,
             status_forcelist=[429, 500, 502, 503, 504],
             allowed_methods=["POST"],
-        )
+        )  # type: ignore
     )
     session.mount("http://", adapter)
     session.mount("https://", adapter)

diff --git a/python/cog/types.py b/python/cog/types.py
@@ -8,8 +8,8 @@
 import urllib.request
 from typing import Any, Dict, Iterator, List, Optional, TypeVar, Union
 
-import requests
-from pydantic import Field, SecretStr
+from ._vendor import requests
+from ._vendor.pydantic import Field, SecretStr
 
 FILENAME_ILLEGAL_CHARS = set("\u0000/")