Windows readline fix, plus run CI against macOS and Windows

* Run CI on Windows and macOS as well as Ubuntu, refs #407
* Use pyreadline3 on win32
* Back to fail-fast since we have a bigger matrix now
* Mark some tests as xfail on windows

.github/workflows/test.yml

@@ -7,16 +7,16 @@ permissions:
 
 jobs:
   test:
-    runs-on: ubuntu-latest
+    runs-on: ${{ matrix.os }}
     strategy:
-      fail-fast: false
       matrix:
+        os: [ubuntu-latest, macos-latest, windows-latest]
         python-version: ["3.8", "3.9", "3.10", "3.11", "3.12"]
         pydantic: ["==1.10.2", ">=2.0.0"]
     steps:
-    - uses: actions/checkout@v3
+    - uses: actions/checkout@v4
     - name: Set up Python ${{ matrix.python-version }}
-      uses: actions/setup-python@v4
+      uses: actions/setup-python@v5
       with:
         python-version: ${{ matrix.python-version }}
         cache: pip
@@ -29,20 +29,25 @@ jobs:
       run: |
         pytest
     - name: Check if cog needs to be run
+      if: matrix.os != 'windows-latest'
       run: |
         cog --check \
           -p "import sys, os; sys._called_from_test=True; os.environ['LLM_USER_PATH'] = '/tmp'" \
           docs/**/*.md docs/*.md
     - name: Run Black
+      if: matrix.os != 'windows-latest'
       run: |
         black --check .
     - name: Run mypy
+      if: matrix.os != 'windows-latest'
       run: |
         mypy llm
     - name: Run ruff
+      if: matrix.os != 'windows-latest'
       run: |
         ruff .
     - name: Run test-llm-load-plugins.sh
+      if: matrix.os != 'windows-latest'
       run: |
         llm install llm-cluster llm-mistral
         ./tests/test-llm-load-plugins.sh

docs/changelog.md

@@ -1,5 +1,10 @@
 # Changelog
 
+(v0_13_1)=
+## 0.13.1 (2024-01-26)
+
+- Fix for `No module named 'readline'` error on Windows. [#407](https://github.com/simonw/llm/issues/407)
+
 (v0_13)=
 ## 0.13 (2024-01-26)
 

setup.py

@@ -1,7 +1,7 @@
 from setuptools import setup, find_packages
 import os
 
-VERSION = "0.13"
+VERSION = "0.13.1"
 
 
 def get_long_description():
@@ -47,6 +47,7 @@ def get_long_description():
         "python-ulid",
         "setuptools",
         "pip",
+        "pyreadline3; sys_platform == 'win32'",
     ],
     extras_require={
         "test": [

tests/test_chat.py

@@ -2,6 +2,7 @@
 import llm.cli
 from unittest.mock import ANY
 import pytest
+import sys
 
 
 def test_mock_model(mock_model):
@@ -16,6 +17,7 @@ def test_mock_model(mock_model):
     assert response2.text() == "second"
 
 
+@pytest.mark.xfail(sys.platform == "win32", reason="Expected to fail on Windows")
 def test_chat_basic(mock_model, logs_db):
     runner = CliRunner()
     mock_model.enqueue(["one world"])
@@ -112,6 +114,7 @@ def test_chat_basic(mock_model, logs_db):
     ]
 
 
+@pytest.mark.xfail(sys.platform == "win32", reason="Expected to fail on Windows")
 def test_chat_system(mock_model, logs_db):
     runner = CliRunner()
     mock_model.enqueue(["I am mean"])
@@ -148,6 +151,7 @@ def test_chat_system(mock_model, logs_db):
     ]
 
 
+@pytest.mark.xfail(sys.platform == "win32", reason="Expected to fail on Windows")
 def test_chat_options(mock_model, logs_db):
     runner = CliRunner()
     mock_model.enqueue(["Some text"])
@@ -175,6 +179,7 @@ def test_chat_options(mock_model, logs_db):
     ]
 
 
+@pytest.mark.xfail(sys.platform == "win32", reason="Expected to fail on Windows")
 @pytest.mark.parametrize(
     "input,expected",
     (

tests/test_embed_cli.py

@@ -5,6 +5,7 @@
 import pathlib
 import pytest
 import sqlite_utils
+import sys
 from unittest.mock import ANY
 
 
@@ -422,6 +423,7 @@ def multi_files(tmpdir):
     return db_path, tmpdir / "files"
 
 
+@pytest.mark.xfail(sys.platform == "win32", reason="Expected to fail on Windows")
 @pytest.mark.parametrize("scenario", ("single", "multi"))
 def test_embed_multi_files(multi_files, scenario):
     db_path, files = multi_files

tests/test_keys.py

@@ -3,8 +3,10 @@
 from llm.cli import cli
 import pathlib
 import pytest
+import sys
 
 
+@pytest.mark.xfail(sys.platform == "win32", reason="Expected to fail on Windows")
 @pytest.mark.parametrize("env", ({}, {"LLM_USER_PATH": "/tmp/llm-keys-test"}))
 def test_keys_in_user_path(monkeypatch, env, user_path):
     for key, value in env.items():
@@ -19,6 +21,7 @@ def test_keys_in_user_path(monkeypatch, env, user_path):
     assert result.output.strip() == expected
 
 
+@pytest.mark.xfail(sys.platform == "win32", reason="Expected to fail on Windows")
 def test_keys_set(monkeypatch, tmpdir):
     user_path = tmpdir / "user/keys"
     monkeypatch.setenv("LLM_USER_PATH", str(user_path))

tests/test_llm.py

@@ -8,6 +8,7 @@
 import pytest
 import re
 import sqlite_utils
+import sys
 from ulid import ULID
 from unittest import mock
 
@@ -96,6 +97,7 @@ def test_logs_json(n, log_path):
     assert len(logs) == expected_length
 
 
+@pytest.mark.xfail(sys.platform == "win32", reason="Expected to fail on Windows")
 @pytest.mark.parametrize("env", ({}, {"LLM_USER_PATH": "/tmp/llm-user-path"}))
 def test_logs_path(monkeypatch, env, user_path):
     for key, value in env.items():