New data-retrieval logic for tasks page

Author: tcompa

.github/workflows/task_list.yaml

@@ -0,0 +1,31 @@
+name: Task-list page
+
+on:
+  push:
+    branches: ["main"]
+  pull_request:
+    branches: ["main"]
+  workflow_dispatch:
+
+
+jobs:
+  retrieve-tasks-data:
+    runs-on: ubuntu-latest
+  
+    steps:
+      - uses: actions/checkout@v4
+        with:
+          fetch-depth: 1
+
+      - uses: actions/setup-python@v5
+        with:
+          python-version: 3.11
+          cache: pip
+
+      - name: Install requirements
+        run: python3 -m pip install -r tasks/data_retrieval/requirements.txt
+
+      - name: Fetch tasks data
+        run: python3 -u tasks/data_retrieval/create_tasks_data.py
+
+      - run: cat tasks/data_retrieval/tasks_data.json

tasks/data_retrieval/.gitignore

@@ -0,0 +1,3 @@
+downloads
+venv
+tasks_data.json

tasks/data_retrieval/create_tasks_data.py

@@ -0,0 +1,244 @@
+import time
+from pathlib import Path
+import requests
+import json
+from os import environ
+from zipfile import ZipFile
+from typing import Any
+from pydantic import BaseModel
+from typing import Optional, Literal
+
+
+DOWNLOAD_FOLDER = Path(__file__).parent / "downloads"
+DOWNLOAD_FOLDER.mkdir(exist_ok=True)
+
+
+class TaskReadV2(BaseModel):
+    """
+    Based on
+    https://github.com/fractal-analytics-platform/fractal-server/blob/main/fractal_server/app/schemas/v2/task.py
+    """
+
+    name: str
+    type: Literal["parallel", "non_parallel", "compound"]
+    source: Optional[str] = None
+    version: Optional[str] = None
+    docs_info: Optional[str] = None
+    docs_link: Optional[str] = None
+    input_types: dict[str, bool]
+    output_types: dict[str, bool]
+    category: Optional[str] = None
+    modality: Optional[str] = None
+    authors: Optional[str] = None
+    tags: list[str]
+
+    class Config:
+        extra = "forbid"
+
+
+def parse_wheel_filename(wheel_path: str) -> dict[str, str]:
+    """
+    Given a wheel-file name or path, extract distribution and version.
+    """
+    wheel_filename = wheel_path.split("/")[-1]
+    parts = wheel_filename.split("-")
+    return dict(name=parts[0], version=parts[1])
+
+
+def load_manifest_from_zip(wheel_path: str) -> dict[str, Any]:
+    """
+    Given a wheel file on-disk, extract the Fractal manifest.
+    """
+    with ZipFile(wheel_path) as wheel:
+        namelist = wheel.namelist()
+        try:
+            manifest = next(
+                name for name in namelist if "__FRACTAL_MANIFEST__.json" in name
+            )
+        except StopIteration:
+            msg = f"{wheel_path} does not include __FRACTAL_MANIFEST__.json"
+            raise ValueError(msg)
+        with wheel.open(manifest) as manifest_fd:
+            manifest_dict = json.load(manifest_fd)
+    return manifest_dict
+
+
+def download_file(url: str) -> str:
+    file_name = url.split("/")[-1]
+    response = requests.get(url, stream=True)
+    file_path = (DOWNLOAD_FOLDER / file_name).as_posix()
+    with open(file_path, "wb") as f:
+        for data in response.iter_content():
+            f.write(data)
+    return file_path
+
+
+def handle_pypi_project(pypi_project_url: str) -> dict[str, Any]:
+    """
+    Example: https://pypi.org/project/fractal-tasks-core
+    """
+
+    # Extract project_name
+    parts = pypi_project_url.split("/")
+    if parts[:4] != ["https:", "", "pypi.org", "project"]:
+        raise ValueError(
+            f"Invalid {pypi_project_url=}.\n"
+            "Valid example: https://pypi.org/project/fractal-tasks-core"
+        )
+    project_name = parts[4]
+
+    # Fetch and parse PyPI information
+    pypi_api_url = f"https://pypi.org/pypi/{project_name}/json"
+    res = requests.get(pypi_api_url)
+    response_data = res.json()
+    if not res.status_code == 200:
+        raise RuntimeError(f"Invalid response from {pypi_api_url}: {res}")
+    latest_version = response_data["info"]["version"]
+    releases = response_data["releases"]
+    latest_release = releases[latest_version]
+    latest_release_wheel_assets = [
+        item for item in latest_release if item["filename"].endswith(".whl")
+    ]
+    if len(latest_release_wheel_assets) > 1:
+        raise ValueError(
+            f"Found more than one wheel asset in release {latest_version}."
+        )
+    latest_release_wheel_asset = latest_release_wheel_assets[0]
+    latest_release_wheel_asset_url = latest_release_wheel_asset["url"]
+
+    # Download wheel and parse manifest
+    wheel_path = download_file(latest_release_wheel_asset_url)
+    info = parse_wheel_filename(wheel_path)
+    manifest = load_manifest_from_zip(wheel_path)
+    Path(wheel_path).unlink()
+
+    return dict(manifest=manifest, **info)
+
+
+def handle_github_repository(github_url: str) -> dict[str, Any]:
+    """
+    Example:
+    https://github.com/fractal-analytics-platform/fractal-lif-converters/
+    """
+
+    # Extract owner and repository
+    parts = github_url.split("/")
+    if parts[:3] != ["https:", "", "github.com"]:
+        print(parts)
+        raise ValueError(
+            f"Invalid {github_url=}.\n"
+            "Valid example: https://github.com/fractal-analytics-platform/fractal-lif-converters"
+        )
+    owner, repository = parts[3:5]
+
+    # Fetch and parse GitHub information
+    github_api_url = (
+        f"https://api.github.com/repos/{owner}/{repository}/releases/latest"
+    )
+    headers = {
+        "Accept": "application/vnd.github+json",
+        "X-GitHub-Api-Version": "2022-11-28",
+    }
+    res = requests.get(github_api_url, headers=headers)
+    if not res.status_code == 200:
+        raise RuntimeError(f"Invalid response from {github_api_url}: {res}")
+    assets = res.json()["assets"]
+    wheel_assets = [asset for asset in assets if asset["name"].endswith(".whl")]
+    if len(wheel_assets) > 1:
+        raise ValueError("Found more than one wheel asset in latest GitHub release.")
+    wheel_asset = wheel_assets[0]
+    wheel_asset_browser_download_url = wheel_asset["browser_download_url"]
+
+    # Download wheel and parse manifest
+    wheel_path = download_file(wheel_asset_browser_download_url)
+    info = parse_wheel_filename(wheel_path)
+    manifest = load_manifest_from_zip(wheel_path)
+    Path(wheel_path).unlink()
+
+    return dict(manifest=manifest, **info)
+
+
+def get_package_info(source: str) -> dict[str, Any]:
+    if source.startswith("https://github.com"):
+        return handle_github_repository(source)
+    elif source.startswith("https://pypi.org"):
+        return handle_pypi_project(source)
+    else:
+        raise ValueError(f"Invalid {source=}.")
+
+
+def _get_task_type(
+    task: dict[str, Any],
+) -> Literal["parallel", "non_parallel", "compound"]:
+    np = task.get("executable_non_parallel", None)
+    p = task.get("executable_parallel", None)
+    if p and np:
+        return "compound"
+    elif p and not np:
+        return "parallel"
+    elif np and not p:
+        return "non_parallel"
+    else:
+        raise ValueError(f"Invalid task with {p=} and {np=}.")
+
+
+COLUMN_NAMES = [
+    "version",
+    "name",
+    "category",
+    "modality",
+    "tags",
+    "input_types",
+    "output_types",
+    "docs_link",
+]
+COLUMN_DEFAULTS = {
+    "input_types": {},
+    "output_types": {},
+    "tags": [],
+}
+COLUMN_TITLES = list(map(str.title, COLUMN_NAMES))
+
+
+# Read and filter list of sources
+sources_file = Path(__file__).parent / "sources.txt"
+with sources_file.open("r") as f:
+    sources = f.read().splitlines()
+sources = [
+    source
+    for source in sources
+    if not (source.startswith("#") or source == "")
+]
+
+TASKS = []
+for source in sources:
+    t_start = time.perf_counter()
+    print(f"START processing {source=}")
+    try:
+        new_tasks = []
+        data = get_package_info(source)
+        pkg_name = data["name"]
+        pkg_version = data.get("version")
+        pkg_task_list = data["manifest"]["task_list"]
+        for task in pkg_task_list:
+            new_task = dict()
+            for column_name in COLUMN_NAMES:
+                new_task[column_name] = task.get(
+                    column_name, COLUMN_DEFAULTS.get(column_name, None)
+                )
+            new_task["version"] = pkg_version
+            new_task["type"] = _get_task_type(task)
+            TaskReadV2(**new_task)
+            new_tasks.append(new_task)
+    except Exception as e:
+        print(f"ERROR, skip.\nOriginal error:\n{str(e)}")
+    TASKS.extend(new_tasks)
+    t_end = time.perf_counter()
+    print(f"END processing {source=} - elapsed {t_end-t_start:.3f} s.")
+    print()
+
+output_file = Path(__file__).parent / "tasks_data.json"
+with output_file.open("w") as f:
+    json.dump(TASKS, f, indent=2)
+
+DOWNLOAD_FOLDER.rmdir()

tasks/data_retrieval/requirements.txt

@@ -0,0 +1,2 @@
+requests
+pydantic

tasks/data_retrieval/sources.txt

@@ -0,0 +1,16 @@
+# PyPI
+https://pypi.org/project/fractal-tasks-core/
+https://pypi.org/project/fractal-faim-ipa
+https://pypi.org/project/fractal-lif-converters
+https://pypi.org/project/operetta-compose
+
+# GitHub releases with wheels
+https://github.com/fractal-analytics-platform/fractal-lif-converters/
+
+
+# https://github.com/fractal-analytics-platform/fractal-helper-tasks
+# https://github.com/fmi-basel/gliberal-scMultipleX
+# https://github.com/Apricot-Therapeutics/APx_fractal_task_collection
+# https://github.com/fractal-analytics-platform/fractal-plantseg-tasks
+# https://github.com/m-albert/fractal-ome-zarr-hcs-stitching/archive
+# https://github.com/fractal-analytics-platform/fractal-ilastik-tasksC/archive/refs/tags/0.1.1.zip