Add GHA action (#7)

* add gha action an

* publish index files

* add checkout

* use github.token

* remove this checkout?

* output to temp

* try explicit working directories

* try github.action_path

* use default env

* change yaml str type

* try with github.action_path again

* update deps

* try --manifest-path

* remove quotes

* a bit more verbosity in deployment decision

* typo

* always add noarch

* Disable GH Pages on tests

.github/workflows/test.yml

@@ -54,3 +54,18 @@ jobs:
       - name: Build recipe (${{ env.PIXI_ENV_NAME }})
         if: matrix.python-version == '310'
         run: pixi run build
+
+  action:
+    runs-on: ubuntu-latest
+    permissions:
+      contents: write  # to deploy to GH Pages automatically
+    steps:
+      - name: Checkout
+        uses: actions/checkout@44c2b7a8a4ea60a981eaca3cf939b5f4305c123b # v4.1.5
+        with:
+          fetch-depth: 0
+      - uses: ./
+        with:
+          channel: conda-forge
+          keep-trees: python=3.9
+          gh-pages-branch: ''  # disable on main

README.md

@@ -1,3 +1,36 @@
 # conda-subchannel
 
 Create subsets of conda channels thanks to CEP-15 metadata
+
+## conda plugin
+
+```bash
+$ conda install -n base conda-subchannel
+$ conda subchannel --channel=conda-forge python=3.9
+$ python -m http.serve --directory subchannel/
+```
+
+## Github Actions action
+
+```yaml
+name: Create conda subchannel
+
+on:
+  push:
+    branches:
+      - main
+  pull_request:
+
+jobs:
+  build-and-deploy:
+    runs-on: ubuntu-latest
+    permissions:
+      contents: write  # to deploy to GH Pages automatically
+    concurrency:
+      group: ${{ github.workflow }}-${{ github.ref }}
+    steps:
+      - uses: jaimergp/conda-subchannel@main
+        with:
+          channel: conda-forge
+          keep-trees: python=3.9
+```

action.yml

@@ -0,0 +1,124 @@
+name: 'conda subchannel'
+description: 'Republish a subset of an existing conda channel'
+inputs:
+  channel:
+    description: "Source conda channel"
+    required: true
+  repodata-fn:
+    description: "Source repodata file to process from channel"
+    required: false
+    default: "repodata.json"
+  subdirs:
+    description: "List of platforms to support, space separated. Defaults to linux-64. Noarch is always included" 
+    required: false
+    default: "linux-64"
+  after:
+    description: "Timestamp as ts:<float> or date as YYYY-[MM[-DD[-HH[-MM[-SS]]]]]"
+    required: false
+    default: ""
+  before:
+    description: "Timestamp as ts:<float> or date as YYYY-[MM[-DD[-HH[-MM[-SS]]]]]"
+    required: false
+    default: ""
+  keep-trees:
+    description: "Keep packages matching these specs and their dependencies. Space separated"
+    required: false
+    default: ""
+  keep-specs:
+    description: "Keep packages matching these specs only. Space separated"
+    required: false
+    default: ""
+  remove-specs:
+    description: "Remove packages matching these specs. Space separated"
+    required: false
+    default: ""
+  gh-pages-branch:
+    description: "Name of the branch for the GH Pages deployment. Set to `''` to disable."
+    required: false
+    default: gh-pages
+outputs:
+  output-directory:
+    description: "Path to the directory containing the subchannel data"
+    value: ${{ steps.validate.outputs.output-path }}
+runs:
+  using: "composite"
+  steps:
+    - name: Check Runner OS
+      if: ${{ runner.os != 'Linux' }}
+      shell: bash
+      run: |
+        echo "::error title=⛔ error hint::Only Linux is supported"
+        exit 1
+    - name: Validate arguments
+      shell: bash
+      id: validate
+      run: |
+        if [[ 
+          "${{ inputs.after }}" == "" 
+          && "${{ inputs.before }}" == "" 
+          && "${{ inputs.keep-trees }}" == "" 
+          && "${{ inputs.keep-specs }}" == "" 
+          && "${{ inputs.remove-specs }}" == "" 
+        ]]; then
+          echo "::error title=⛔ error hint::At least one of `after`, `before`, `keep-trees`, `keep-specs` or `remove-specs` must be set"
+          exit 1
+        fi
+        mkdir -p "${{ runner.temp }}/subchannel"
+        echo "output-directory=${{ runner.temp }}/subchannel" >> $GITHUB_OUTPUT
+    - uses: prefix-dev/setup-pixi@632d17935141ec801697e2c359784b878adecbbe # v0.6.0
+      with:
+        environments: default
+        manifest-path: ${{ github.action_path }}/pyproject.toml
+    - name: Setup project
+      shell: bash
+      run: cd "${{ github.action_path }}" && pixi run --environment default dev
+    - name: Run subchannel
+      shell: pixi run --manifest-path "${{ github.action_path }}/pyproject.toml" --environment default bash -e {0}
+      run: |
+        args="--channel ${{ inputs.channel }}"
+        args+=" --output ${{ steps.validate.outputs.output-directory }}"
+        args+=" --repodata-fn ${{ inputs.repodata-fn }}"
+        for subdir in ${{ inputs.subdirs }}; do
+          args+=" --subdir $subdir"
+        done
+        if [[ "${{ inputs.after }}" != "" ]]; then
+          args+=" --after ${{ inputs.after }}"
+        fi
+        if [[ "${{ inputs.before }}" != "" ]]; then
+          args+=" --before ${{ inputs.before }}"
+        fi
+        if [[ "${{ inputs.keep-trees }}" != "" ]]; then
+          for spec in ${{ inputs.keep-trees }}; do
+            args+=" --keep-tree $spec"
+          done
+        fi
+        if [[ "${{ inputs.keep-specs }}" != "" ]]; then
+          for spec in ${{ inputs.keep-specs }}; do
+            args+=" --keep $spec"
+          done
+        fi
+        if [[ "${{ inputs.remove-specs }}" != "" ]]; then
+          for spec in ${{ inputs.remove-specs }}; do
+            args+=" --remove $spec"
+          done
+        fi
+        echo "Running: conda subchannel $args"
+        conda subchannel $args
+    - name: Decide deployment
+      id: decide
+      shell: bash
+      run: |
+        if [[ "${{ inputs.gh-pages-branch }}" != "" && "${{ github.ref }}" == "refs/heads/main" ]]; then
+          echo "deploy=true" >> $GITHUB_OUTPUT
+        else
+          echo "Will skip deployment to GH Pages."
+          echo "deploy=false" >> $GITHUB_OUTPUT
+        fi
+    - uses: peaceiris/actions-gh-pages@4f9cc6602d3f66b9c108549d475ec49e8ef4d45e # v4.0.0
+      if: steps.decide.outputs.deploy == 'true'
+      with:
+        github_token: ${{ github.token }}
+        publish_branch: ${{ inputs.gh-pages-branch }}
+        publish_dir: ${{ steps.validate.outputs.output-directory }}
+        user_name: 'github-actions[bot]'
+        user_email: 'github-actions[bot]@users.noreply.github.com'

conda_subchannel/cli.py

@@ -35,7 +35,13 @@ def date_argument(date: str) -> float:
 
 
 def configure_parser(parser: argparse.ArgumentParser):
-    parser.add_argument("-c", "--channel", required=True, dest="channel")
+    parser.add_argument(
+        "-c",
+        "--channel",
+        required=True,
+        dest="channel",
+        help="Source conda channel.",
+    )
     parser.add_argument(
         "--repodata-fn",
         default=REPODATA_FN,
@@ -93,9 +99,10 @@ def execute(args: argparse.Namespace) -> int:
         raise ArgumentError("Please provide at least one filter.")
 
     with Spinner("Syncing source channel"):
-        subdir_datas = _fetch_channel(
-            args.channel, args.subdirs or context.subdirs, args.repodata_fn
-        )
+        subdirs = args.subdirs or context.subdirs
+        if "noarch" not in subdirs:
+            subdirs = *subdirs, "noarch"
+        subdir_datas = _fetch_channel(args.channel, subdirs, args.repodata_fn)
     for sd in sorted(subdir_datas, key=lambda sd: sd.channel.name):
         print(" -", sd.channel.name, sd.channel.subdir)
 
@@ -117,6 +124,6 @@ def execute(args: argparse.Namespace) -> int:
 
     with Spinner(f"Writing output to {args.output}"):
         repodatas = _dump_records(records, args.channel)
-        _write_to_disk(repodatas, args.output)
+        _write_to_disk(args.channel, repodatas, args.output)
 
     return 0

conda_subchannel/core.py

@@ -1,8 +1,10 @@
 from __future__ import annotations
 
 import bz2
+import hashlib
 import json
 import logging
+from datetime import datetime, timezone
 from pathlib import Path
 from typing import TYPE_CHECKING
 
@@ -162,7 +164,55 @@ def _dump_records(
     return repodatas
 
 
+def _checksum(path, algorithm, buffersize=65536):
+    hash_impl = getattr(hashlib, algorithm)()
+    with open(path, "rb") as f:
+        for block in iter(lambda: f.read(buffersize), b""):
+            hash_impl.update(block)
+    return hash_impl.hexdigest()
+
+
+def _write_channel_index_md(source_channel: Channel, channel_path: Path):
+    channel_path = Path(channel_path)
+    lines = [
+        f"# {channel_path.name}",
+        "",
+        f"Derived from [{source_channel.name}]({source_channel.base_url})",
+        "",
+        ""
+    ]
+    for subdir in channel_path.glob("*"):
+        if subdir.is_file():
+            continue
+        lines[-1] += f"[{subdir.name}]({subdir.name}) "
+    (channel_path / "index.md").write_text("\n".join(lines))
+
+
+def _write_subdir_index_md(subdir_path: Path):
+    subdir_path = Path(subdir_path)
+    lines = [
+        f"# {'/'.join(subdir_path.parts[-2:])}",
+        "| Filename | Size (B) | Last modified | SHA256 | MD5 |",
+        "|----------|----------|---------------|--------|-----|",
+    ]
+    for path in sorted(subdir_path.glob("*")):
+        if path.name == "index.md":
+            continue
+        stat = path.stat()
+        size = stat.st_size
+        lastmod = datetime.fromtimestamp(stat.st_mtime)
+        sha256 = _checksum(path, "sha256")
+        md5 = _checksum(path, "md5")
+        lines.append(
+            f"| [{path.name}]({path.name}) | {size} | {lastmod} | `{sha256}` | `{md5}` |"
+        )
+    lines.append("")
+    lines.append(f"> Last modified on {datetime.now(tz=timezone.utc)}")
+    (subdir_path / "index.md").write_text("\n".join(lines))
+
+
 def _write_to_disk(
+    source_channel: Channel | str,
     repodatas: dict[str, dict[str, Any]],
     path: os.PathLike | str,
     outputs: Iterable[str] = ("bz2", "zstd"),
@@ -185,8 +235,13 @@ def _write_to_disk(
                     level=ZSTD_COMPRESS_LEVEL, threads=ZSTD_COMPRESS_THREADS
                 ).compress(json_contents.encode("utf-8"))
                 fo.write(repodata_zst_content)
+        _write_subdir_index_md(path / subdir)
+
     # noarch must always be present
     noarch_repodata = path / "noarch" / "repodata.json"
     if not noarch_repodata.is_file():
         noarch_repodata.parent.mkdir(parents=True, exist_ok=True)
         noarch_repodata.write_text("{}")
+        _write_subdir_index_md(path / "noarch")
+
+    _write_channel_index_md(Channel(source_channel), path)