MAINT Store data in bytes not io.BytesIO (#91)

* Store data in bytes instead of io

* Fix lint

* Fix typo

* Address reviews

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

* Fix test package config

* Use normalized name

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

---------

Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>

micropip/_compat_in_pyodide.py

@@ -1,5 +1,4 @@
-from io import BytesIO
-from typing import IO
+from pathlib import Path
 from urllib.parse import urlparse
 
 from pyodide._package_loader import get_dynlibs
@@ -20,15 +19,14 @@
     # Otherwise, this is pytest test collection so let it go.
 
 
-async def fetch_bytes(url: str, kwargs: dict[str, str]) -> IO[bytes]:
+async def fetch_bytes(url: str, kwargs: dict[str, str]) -> bytes:
     parsed_url = urlparse(url)
     if parsed_url.scheme == "emfs":
-        return open(parsed_url.path, "rb")
+        return Path(parsed_url.path).read_bytes()
     if parsed_url.scheme == "file":
-        result_bytes = (await loadBinaryFile(parsed_url.path)).to_bytes()
-    else:
-        result_bytes = await (await pyfetch(url, **kwargs)).bytes()
-    return BytesIO(result_bytes)
+        return (await loadBinaryFile(parsed_url.path)).to_bytes()
+
+    return await (await pyfetch(url, **kwargs)).bytes()
 
 
 async def fetch_string_and_headers(

micropip/_compat_not_in_pyodide.py

@@ -1,5 +1,4 @@
 import re
-from io import BytesIO
 from pathlib import Path
 from typing import IO, Any
 
@@ -20,9 +19,8 @@ def _fetch(url: str, kwargs: dict[str, Any]) -> addinfourl:
     return urlopen(Request(url, **kwargs))
 
 
-async def fetch_bytes(url: str, kwargs: dict[str, Any]) -> IO[bytes]:
-    response = _fetch(url, kwargs=kwargs)
-    return BytesIO(response.read())
+async def fetch_bytes(url: str, kwargs: dict[str, Any]) -> bytes:
+    return _fetch(url, kwargs=kwargs).read()
 
 
 async def fetch_string_and_headers(

micropip/wheelinfo.py

@@ -1,10 +1,11 @@
 import asyncio
 import hashlib
+import io
 import json
 import zipfile
 from dataclasses import dataclass
 from pathlib import Path
-from typing import IO, Any
+from typing import Any
 from urllib.parse import ParseResult, urlparse
 
 from packaging.requirements import Requirement
@@ -39,7 +40,7 @@ class WheelInfo:
 
     # Fields below are only available after downloading the wheel, i.e. after calling `download()`.
 
-    _data: IO[bytes] | None = None  # Wheel file contents.
+    _data: bytes | None = None  # Wheel file contents.
     _metadata: Metadata | None = None  # Wheel metadata.
     _requires: list[Requirement] | None = None  # List of requirements.
 
@@ -109,7 +110,7 @@ async def install(self, target: Path) -> None:
             raise RuntimeError(
                 "Micropip internal error: attempted to install wheel before downloading it?"
             )
-        self._validate()
+        _validate_sha256_checksum(self._data, self.sha256)
         self._extract(target)
         await self._load_libraries(target)
         self._set_installer()
@@ -119,7 +120,7 @@ async def download(self, fetch_kwargs: dict[str, Any]):
             return
 
         self._data = await self._fetch_bytes(fetch_kwargs)
-        with zipfile.ZipFile(self._data) as zf:
+        with zipfile.ZipFile(io.BytesIO(self._data)) as zf:
             metadata_path = wheel_dist_info_dir(zf, self.name) + "/" + Metadata.PKG_INFO
             self._metadata = Metadata(zipfile.Path(zf, metadata_path))
 
@@ -153,20 +154,9 @@ async def _fetch_bytes(self, fetch_kwargs: dict[str, Any]):
                     "Check if the server is sending the correct 'Access-Control-Allow-Origin' header."
                 ) from e
 
-    def _validate(self):
-        if self.sha256 is None:
-            # No checksums available, e.g. because installing
-            # from a different location than PyPI.
-            return
-
-        assert self._data
-        sha256_actual = _generate_package_hash(self._data)
-        if sha256_actual != self.sha256:
-            raise ValueError("Contents don't match hash")
-
     def _extract(self, target: Path) -> None:
         assert self._data
-        with zipfile.ZipFile(self._data) as zf:
+        with zipfile.ZipFile(io.BytesIO(self._data)) as zf:
             zf.extractall(target)
             self._dist_info = target / wheel_dist_info_dir(zf, self.name)
 
@@ -198,16 +188,20 @@ async def _load_libraries(self, target: Path) -> None:
         TODO: integrate with pyodide's dynamic library loading mechanism.
         """
         assert self._data
-        dynlibs = get_dynlibs(self._data, ".whl", target)
+        dynlibs = get_dynlibs(io.BytesIO(self._data), ".whl", target)
         await asyncio.gather(*map(lambda dynlib: loadDynlib(dynlib, False), dynlibs))
 
 
-def _generate_package_hash(data: IO[bytes]) -> str:
-    """
-    Generate a SHA256 hash of the package data.
-    """
-    sha256_hash = hashlib.sha256()
-    data.seek(0)
-    while chunk := data.read(4096):
-        sha256_hash.update(chunk)
-    return sha256_hash.hexdigest()
+def _validate_sha256_checksum(data: bytes, expected: str | None = None) -> None:
+    if expected is None:
+        # No checksums available, e.g. because installing
+        # from a different location than PyPI.
+        return
+
+    actual = _generate_package_hash(data)
+    if actual != expected:
+        raise RuntimeError(f"Invalid checksum: expected {expected}, got {actual}")
+
+
+def _generate_package_hash(data: bytes) -> str:
+    return hashlib.sha256(data).hexdigest()

tests/conftest.py

@@ -257,7 +257,7 @@ def write_file(filename, contents):
 
         tmp.seek(0)
 
-        return tmp
+        return tmp.read()
 
 
 @pytest.fixture

tests/test_data/test_wheel_uninstall/pyproject.toml

@@ -1,11 +1,11 @@
 [project]
-name = "test_wheel_uninstall"
+name = "test-wheel-uninstall"
 description = "Test wheel uninstall"
 requires-python = ">=3.10"
 version = "1.0.0"
 
 [tool.setuptools]
-packages = ["deep", "deep.deep", "shallow", "test_wheel_uninstall"]
+packages = ["deep", "deep.deep", "shallow", "test_wheel_uninstall", "deep.data"]
 py-modules = ["top_level"]
 
 [tool.setuptools.package-data]

tests/test_install.py

@@ -370,8 +370,6 @@ async def run_test(selenium, url, name, version):
 
 @pytest.mark.asyncio
 async def test_custom_index_urls(mock_package_index_json_api, monkeypatch):
-    from io import BytesIO
-
     mock_server_fake_package = mock_package_index_json_api(
         pkgs=["fake-pkg-micropip-test"]
     )
@@ -381,7 +379,7 @@ async def test_custom_index_urls(mock_package_index_json_api, monkeypatch):
     async def _mock_fetch_bytes(url, *args):
         nonlocal _wheel_url
         _wheel_url = url
-        return BytesIO(b"fake wheel")
+        return b"fake wheel"
 
     from micropip import wheelinfo
 

tests/test_uninstall.py

@@ -3,9 +3,10 @@
 import pytest
 from pytest_pyodide import run_in_pyodide, spawn_web_server
 from conftest import SNOWBALL_WHEEL, TEST_WHEEL_DIR
-from packaging.utils import parse_wheel_filename
+from packaging.utils import parse_wheel_filename, canonicalize_name
 
 TEST_PACKAGE_NAME = "test_wheel_uninstall"
+TEST_PACKAGE_NAME_NORMALIZED = canonicalize_name(TEST_PACKAGE_NAME)
 
 
 @pytest.fixture(scope="module")
@@ -19,15 +20,15 @@ def test_wheel_url(test_wheel_path):
 
 def test_basic(selenium_standalone_micropip, test_wheel_url):
     @run_in_pyodide()
-    async def run(selenium, pkg_name, wheel_url):
+    async def run(selenium, pkg_name, pkg_name_normalized, wheel_url):
         import importlib.metadata
         import sys
 
         import micropip
 
         await micropip.install(wheel_url)
 
-        assert pkg_name in micropip.list()
+        assert pkg_name_normalized in micropip.list()
         assert pkg_name not in sys.modules
 
         __import__(pkg_name)
@@ -52,7 +53,12 @@ async def run(selenium, pkg_name, wheel_url):
         # 3. Check that the module is not available with micropip.list()
         assert pkg_name not in micropip.list()
 
-    run(selenium_standalone_micropip, TEST_PACKAGE_NAME, test_wheel_url)
+    run(
+        selenium_standalone_micropip,
+        TEST_PACKAGE_NAME,
+        TEST_PACKAGE_NAME_NORMALIZED,
+        test_wheel_url,
+    )
 
 
 def test_files(selenium_standalone_micropip, test_wheel_url):
@@ -61,13 +67,13 @@ def test_files(selenium_standalone_micropip, test_wheel_url):
     """
 
     @run_in_pyodide()
-    async def run(selenium, pkg_name, wheel_url):
+    async def run(selenium, pkg_name, pkg_name_normalized, wheel_url):
         import importlib.metadata
 
         import micropip
 
         await micropip.install(wheel_url)
-        assert pkg_name in micropip.list()
+        assert pkg_name_normalized in micropip.list()
 
         dist = importlib.metadata.distribution(pkg_name)
         files = dist.files
@@ -86,7 +92,12 @@ async def run(selenium, pkg_name, wheel_url):
 
         assert not dist._path.is_dir(), f"{dist._path} still exists after removal"
 
-    run(selenium_standalone_micropip, TEST_PACKAGE_NAME, test_wheel_url)
+    run(
+        selenium_standalone_micropip,
+        TEST_PACKAGE_NAME,
+        TEST_PACKAGE_NAME_NORMALIZED,
+        test_wheel_url,
+    )
 
 
 def test_install_again(selenium_standalone_micropip, test_wheel_url):
@@ -95,20 +106,20 @@ def test_install_again(selenium_standalone_micropip, test_wheel_url):
     """
 
     @run_in_pyodide()
-    async def run(selenium, pkg_name, wheel_url):
+    async def run(selenium, pkg_name, pkg_name_normalized, wheel_url):
         import sys
 
         import micropip
 
         await micropip.install(wheel_url)
 
-        assert pkg_name in micropip.list()
+        assert pkg_name_normalized in micropip.list()
 
         __import__(pkg_name)
 
         micropip.uninstall(pkg_name)
 
-        assert pkg_name not in micropip.list()
+        assert pkg_name_normalized not in micropip.list()
 
         del sys.modules[pkg_name]
 
@@ -121,10 +132,15 @@ async def run(selenium, pkg_name, wheel_url):
 
         await micropip.install(wheel_url)
 
-        assert pkg_name in micropip.list()
+        assert pkg_name_normalized in micropip.list()
         __import__(pkg_name)
 
-    run(selenium_standalone_micropip, TEST_PACKAGE_NAME, test_wheel_url)
+    run(
+        selenium_standalone_micropip,
+        TEST_PACKAGE_NAME,
+        TEST_PACKAGE_NAME_NORMALIZED,
+        test_wheel_url,
+    )
 
 
 def test_warning_not_installed(selenium_standalone_micropip):
@@ -156,7 +172,7 @@ def test_warning_file_removed(selenium_standalone_micropip, test_wheel_url):
     """
 
     @run_in_pyodide()
-    async def run(selenium, pkg_name, wheel_url):
+    async def run(selenium, pkg_name, pkg_name_normalized, wheel_url):
         from importlib.metadata import distribution
         import micropip
         import contextlib
@@ -165,17 +181,17 @@ async def run(selenium, pkg_name, wheel_url):
         with io.StringIO() as buf, contextlib.redirect_stdout(buf):
             await micropip.install(wheel_url)
 
-            assert pkg_name in micropip.list()
+            assert pkg_name_normalized in micropip.list()
 
-            dist = distribution(pkg_name)
+            dist = distribution(pkg_name_normalized)
             files = dist.files
             file1 = files[0]
             file2 = files[1]
 
             file1.locate().unlink()
             file2.locate().unlink()
 
-            micropip.uninstall(pkg_name)
+            micropip.uninstall(pkg_name_normalized)
 
             captured = buf.getvalue()
             logs = captured.strip().split("\n")
@@ -184,7 +200,12 @@ async def run(selenium, pkg_name, wheel_url):
             assert "does not exist" in logs[-1]
             assert "does not exist" in logs[-2]
 
-    run(selenium_standalone_micropip, TEST_PACKAGE_NAME, test_wheel_url)
+    run(
+        selenium_standalone_micropip,
+        TEST_PACKAGE_NAME,
+        TEST_PACKAGE_NAME_NORMALIZED,
+        test_wheel_url,
+    )
 
 
 def test_warning_remaining_file(selenium_standalone_micropip, test_wheel_url):
@@ -193,28 +214,33 @@ def test_warning_remaining_file(selenium_standalone_micropip, test_wheel_url):
     """
 
     @run_in_pyodide()
-    async def run(selenium, pkg_name, wheel_url):
+    async def run(selenium, pkg_name, pkg_name_normalized, wheel_url):
         from importlib.metadata import distribution
         import micropip
         import contextlib
         import io
 
         with io.StringIO() as buf, contextlib.redirect_stdout(buf):
             await micropip.install(wheel_url)
-            assert pkg_name in micropip.list()
+            assert pkg_name_normalized in micropip.list()
 
-            pkg_dir = distribution(pkg_name)._path.parent / "deep"
+            pkg_dir = distribution(pkg_name_normalized)._path.parent / "deep"
             (pkg_dir / "extra-file.txt").touch()
 
-            micropip.uninstall(pkg_name)
+            micropip.uninstall(pkg_name_normalized)
 
             captured = buf.getvalue()
             logs = captured.strip().split("\n")
 
             assert len(logs) == 1
             assert "is not empty after uninstallation" in logs[0]
 
-    run(selenium_standalone_micropip, TEST_PACKAGE_NAME, test_wheel_url)
+    run(
+        selenium_standalone_micropip,
+        TEST_PACKAGE_NAME,
+        TEST_PACKAGE_NAME_NORMALIZED,
+        test_wheel_url,
+    )
 
 
 def test_pyodide_repodata(selenium_standalone_micropip):