CI with GitHub workflows #179:

- support type checks with pyright - add github workflow for lint/typecheck/test - configure inclusion/exclusion rules for black - can now be run from the root dir - move all downloaded models to one folder for caching - adds cache to github workflow - add dependencies for type checking - fix type check issues with latest PyQt5 module - fix download_models minimal missing control nets - ignore windows-only constant in type checks - move image out of LFS for CI tests it's tiny, and the rest of the images aren't needed at the moment
Acly · Dec 5, 2023 · cf127d8 · cf127d8
1 parent 3522e1b
commit cf127d8
Show file tree

Hide file tree

Showing 27 changed files with 159 additions and 81 deletions.
diff --git a/.github/workflows/test.yml b/.github/workflows/test.yml
@@ -0,0 +1,52 @@
+name: Lint, Typecheck and Test
+on: [push, pull_request]
+
+jobs:
+  check:
+    runs-on: ubuntu-latest
+    steps:
+      - name: Checkout
+        uses: actions/checkout@v3
+        with:
+          submodules: true
+      - name: Python
+        uses: actions/setup-python@v4
+        with:
+          python-version: '3.10'
+          cache: 'pip'
+      - name: Dependencies
+        run: pip install -r requirements.txt          
+      - name: Typecheck
+        uses: jakebailey/pyright-action@v1
+        with:
+          pylance-version: latest-release
+      - name: Lint
+        if: ${{ !cancelled() }}
+        uses: psf/black@stable
+
+  test:
+    runs-on: ubuntu-latest
+    steps:
+      - name: Checkout
+        uses: actions/checkout@v3
+        with:
+          submodules: true
+      - name: Python
+        uses: actions/setup-python@v4
+        with:
+          python-version: '3.10'
+          cache: 'pip'
+      - name: Dependencies
+        run: pip install -r requirements.txt
+      - name: Cache models
+        uses: actions/cache@v3
+        with:
+          path: scripts/docker/downloads
+          key: models-v1
+      - name: Download models
+        run: python scripts/download_models.py --minimal scripts/docker/downloads
+      - name: Test installer
+        run: python -m pytest tests/test_server.py -vs --test-install
+      - name: Test
+        run: python -m pytest tests -vs --ci
+
diff --git a/.gitignore b/.gitignore
@@ -7,8 +7,7 @@
 .server
 .vscode
 __pycache__
-scripts/docker/models
-scripts/docker/custom_nodes/*
+scripts/docker/downloads
 tests/.results
 settings.json
 ai_diffusion/styles/*

diff --git a/CONTRIBUTING.md b/CONTRIBUTING.md
@@ -27,7 +27,7 @@ The easiest way to run a development version of the plugin is to use symlinks:
 
 ### Code formatting
 
-The codebase uses [black](https://github.com/psf/black) for formatting. The project root contains a `pyproject.toml` to configure the line length, it should be picked up automatically.
+The codebase uses [black](https://github.com/psf/black) for formatting. You can check locally by running `black` in the repository root, or use an IDE integration.
 
 ### Code style
 
@@ -37,7 +37,9 @@ Code style follows the official Python recommendations. Only exception: no `ALL_
 
 Type annotations should be used where types can't be inferred. Basic type checks are enabled for the project and should not report errors.
 
-The `Krita` module is special in that it is usually only available when running inside Krita. To make type checking work, include `scripts/typeshed` in your `PYTHONPATH`.
+The `Krita` module is special in that it is usually only available when running inside Krita. To make type checking work an interface file is located in `scripts/typeshed`.
+
+You can run `pyright` from the repository root to perform type checks on the entire codebase. This is also done by the CI.
 
 Configuration for VSCode with Pylance (.vscode/settings.json):
 ```
@@ -77,9 +79,9 @@ Everything else has tests. Mostly. If effort is reasonable, tests are expected.
 
 Testing changes to the installer is annoying because of the file sizes involved. There are some things that help. You can preload model files with the following script:
 ```
-python scripts/download_models.py --minimal scripts/docker
+python scripts/download_models.py --minimal scripts/docker/downloads
 ```
-This will download the minimum required models and store them in `scripts/docker` (used as default location because that way the docker build script can use them too).
+This will download the minimum required models and store them in `scripts/docker/downloads` (used as default location because that way the docker build script can use them too).
 
 The following command does some automated testing for installation and upgrade. It starts a local file server which pulls preloaded models, so it's reasonably fast and doesn't download the entire internet.
 ```

diff --git a/ai_diffusion/attention_edit.py b/ai_diffusion/attention_edit.py
@@ -5,7 +5,7 @@
 def select_current_parenthesis_block(
     text: str, cursor_pos: int, open_bracket: str, close_bracket: str
 ) -> Tuple[int, int] | None:
-    """Select the current parenthesis block that the cursor points to. """
+    """Select the current parenthesis block that the cursor points to."""
     # Ensure cursor position is within valid range
     cursor_pos = max(0, min(cursor_pos, len(text)))
 
@@ -50,9 +50,8 @@ def select_current_word(text: str, cursor_pos: int) -> Tuple[int, int]:
 
 def select_on_cursor_pos(text: str, cursor_pos: int) -> Tuple[int, int]:
     """Return a range in the text based on the cursor_position."""
-    return (
-        select_current_parenthesis_block(text, cursor_pos, "(", ")")
-        or select_current_word(text, cursor_pos)
+    return select_current_parenthesis_block(text, cursor_pos, "(", ")") or select_current_word(
+        text, cursor_pos
     )
 
 
@@ -143,4 +142,3 @@ def edit_attention(text: str, positive: bool) -> str:
         if weight == 1.0
         else f"{open_bracket}{attention_string}:{weight:.1f}{close_bracket}"
     )
-
diff --git a/ai_diffusion/client.py b/ai_diffusion/client.py
@@ -128,7 +128,7 @@ class Client:
     lcm_model: dict[SDVersion, str | None]
     supported_sd_versions: list[SDVersion]
     device_info: DeviceInfo
-    nodes_required_inputs: dict[str, dict[str, list[str | list | dict]]] = {}
+    nodes_inputs: dict[str, dict[str, list[str | list | dict]]] = {}
 
     @staticmethod
     async def connect(url=default_url):
@@ -164,7 +164,7 @@ async def connect(url=default_url):
         client.ip_adapter_model = {
             ver: _find_ip_adapter(ip, ver) for ver in [SDVersion.sd15, SDVersion.sdxl]
         }
-        client.nodes_required_inputs["IPAdapterApply"] = nodes["IPAdapterApply"]["input"]["required"]
+        client.nodes_inputs["IPAdapterApply"] = nodes["IPAdapterApply"]["input"]["required"]
 
         # Retrieve upscale models
         client.upscalers = nodes["UpscaleModelLoader"]["input"]["required"]["model_name"][0]
@@ -395,7 +395,7 @@ def _check_workload(self, sdver: SDVersion) -> list[MissingResource]:
 
 def parse_url(url: str):
     url = url.strip("/")
-    url = url.replace('0.0.0.0', '127.0.0.1')
+    url = url.replace("0.0.0.0", "127.0.0.1")
     if not url.startswith("http"):
         url = f"http://{url}"
     return url

diff --git a/ai_diffusion/image.py b/ai_diffusion/image.py
@@ -261,6 +261,7 @@ def make_opaque(self, background=Qt.GlobalColor.white):
     @property
     def data(self):
         ptr = self._qimage.bits()
+        assert ptr is not None, "Accessing data of invalid image"
         ptr.setsize(self._qimage.byteCount())
         return QByteArray(ptr.asstring())
 
@@ -272,7 +273,9 @@ def to_array(self):
         import numpy as np
 
         w, h = self.extent
-        ptr = self._qimage.constBits().asarray(w * h * 4)
+        bits = self._qimage.constBits()
+        assert bits is not None, "Accessing data of invalid image"
+        ptr = bits.asarray(w * h * 4)
         array = np.frombuffer(ptr, np.uint8).reshape(w, h, 4)  # type: ignore
         return array.astype(np.float32) / 255
 

diff --git a/ai_diffusion/network.py b/ai_diffusion/network.py
@@ -164,6 +164,7 @@ async def _try_download(network: QNetworkAccessManager, url: str, path: Path):
         log.info(f"Found {path}.part, resuming download from {out_file.size()} bytes")
         request.setRawHeader(b"Range", f"bytes={out_file.size()}-".encode("utf-8"))
     reply = network.get(request)
+    assert reply is not None, f"Network request for {url} failed: reply is None"
 
     progress_future = asyncio.get_running_loop().create_future()
     finished_future = asyncio.get_running_loop().create_future()

diff --git a/ai_diffusion/server.py b/ai_diffusion/server.py
@@ -17,7 +17,7 @@
 
 
 _exe = ".exe" if is_windows else ""
-_process_flags = subprocess.CREATE_NO_WINDOW if is_windows else 0
+_process_flags = subprocess.CREATE_NO_WINDOW if is_windows else 0  # type: ignore
 
 
 class ServerState(Enum):

diff --git a/ai_diffusion/ui/generation.py b/ai_diffusion/ui/generation.py
@@ -20,6 +20,7 @@
 from ..model import Model
 from ..root import root
 from ..settings import settings
+from ..util import ensure
 from . import theme
 from .widget import (
     WorkspaceSelectWidget,
@@ -94,7 +95,7 @@ def add(self, job: Job):
             self.addItem(item)
 
         scrollbar = self.verticalScrollBar()
-        if scrollbar.isVisible() and scrollbar.value() >= scrollbar.maximum() - 4:
+        if scrollbar and scrollbar.isVisible() and scrollbar.value() >= scrollbar.maximum() - 4:
             self.scrollToBottom()
 
     def update_selection(self):
@@ -118,7 +119,7 @@ def is_finished(self, job: Job):
 
     def prune(self, jobs: JobQueue):
         first_id = next((job.id for job in jobs if self.is_finished(job)), None)
-        while self.count() > 0 and self.item(0).data(Qt.ItemDataRole.UserRole) != first_id:
+        while self.count() > 0 and ensure(self.item(0)).data(Qt.ItemDataRole.UserRole) != first_id:
             self.takeItem(0)
 
     def rebuild(self):
@@ -131,8 +132,9 @@ def item_info(self, item: QListWidgetItem) -> tuple[str, int]:  # job id, image
 
     def handle_preview_click(self, item: QListWidgetItem):
         if item.text() != "" and item.text() != "<no prompt>":
-            prompt = item.data(Qt.ItemDataRole.ToolTipRole)
-            QGuiApplication.clipboard().setText(prompt)
+            if clipboard := QGuiApplication.clipboard():
+                prompt = item.data(Qt.ItemDataRole.ToolTipRole)
+                clipboard.setText(prompt)
 
     def mousePressEvent(self, e: QMouseEvent) -> None:
         # make single click deselect current item (usually requires Ctrl+click)
@@ -151,7 +153,7 @@ def mousePressEvent(self, e: QMouseEvent) -> None:
         return super().mousePressEvent(e)
 
     def _find(self, id: JobQueue.Item):
-        items = (self.item(i) for i in range(self.count()))
+        items = (ensure(self.item(i)) for i in range(self.count()))
         return next((item for item in items if self._item_data(item) == id), None)
 
     def _item_data(self, item: QListWidgetItem):

diff --git a/ai_diffusion/ui/live.py b/ai_diffusion/ui/live.py
@@ -118,7 +118,9 @@ def __init__(self):
 
         self.preview_area = QLabel(self)
         self.preview_area.setSizePolicy(QSizePolicy.Expanding, QSizePolicy.Expanding)
-        self.preview_area.setAlignment(Qt.AlignmentFlag.AlignTop | Qt.AlignmentFlag.AlignLeft)
+        self.preview_area.setAlignment(
+            Qt.AlignmentFlag(Qt.AlignmentFlag.AlignTop | Qt.AlignmentFlag.AlignLeft)
+        )
         layout.addWidget(self.preview_area)
 
     @property

diff --git a/ai_diffusion/ui/settings.py b/ai_diffusion/ui/settings.py
@@ -1055,10 +1055,9 @@ def __init__(self, server: Server):
 
         self.setWindowTitle("Configure Image Diffusion")
         self.setMinimumSize(QSize(840, 480))
-        screen_size = QGuiApplication.primaryScreen().availableSize()
-        self.resize(
-            QSize(max(900, int(screen_size.width() * 0.6)), int(screen_size.height() * 0.8))
-        )
+        if screen := QGuiApplication.primaryScreen():
+            size = screen.availableSize()
+            self.resize(QSize(max(900, int(size.width() * 0.6)), int(size.height() * 0.8)))
 
         layout = QHBoxLayout()
         self.setLayout(layout)

diff --git a/ai_diffusion/ui/theme.py b/ai_diffusion/ui/theme.py
@@ -40,7 +40,9 @@ def sd_version_icon(version: SDVersion, client: Client | None = None):
         return icon("sd-version-15")
     elif version is SDVersion.sdxl:
         return icon("sd-version-xl")
-    return None
+    else:
+        util.client_logger.warning(f"Unresolved SD version {version}, cannot fetch icon")
+        return icon("warning")
 
 
 def logo():

diff --git a/ai_diffusion/ui/widget.py b/ai_diffusion/ui/widget.py
@@ -22,17 +22,17 @@
 import krita
 
 from ..style import Style, Styles
-from ..image import Bounds
 from ..resources import ControlMode
 from ..root import root
 from ..client import filter_supported_styles, resolve_sd_version
 from ..properties import Binding, bind, bind_combo
-from ..jobs import Job, JobKind, JobState, JobQueue
+from ..jobs import JobState, JobQueue
 from ..model import Model, Workspace, ControlLayer
+from ..attention_edit import edit_attention, select_on_cursor_pos
+from ..util import ensure
 from .settings import SettingsDialog
 from .theme import SignalBlocker
 from . import actions, theme
-from ..attention_edit import edit_attention, select_on_cursor_pos
 
 
 class QueueWidget(QToolButton):
@@ -400,7 +400,7 @@ def line_count(self):
     @line_count.setter
     def line_count(self, value: int):
         self._line_count = value
-        fm = QFontMetrics(self.document().defaultFont())
+        fm = QFontMetrics(ensure(self.document()).defaultFont())
         self.setFixedHeight(fm.lineSpacing() * value + 6)
 
     def hasSelectedText(self) -> bool:

diff --git a/ai_diffusion/workflow.py b/ai_diffusion/workflow.py
@@ -422,7 +422,7 @@ def generate(
     sampler_params = _sampler_params(style, live=live)
     batch = 1 if live.is_active else batch
 
-    w = ComfyWorkflow(comfy.nodes_required_inputs)
+    w = ComfyWorkflow(comfy.nodes_inputs)
     model, clip, vae = load_model_with_lora(w, comfy, style, is_live=live.is_active)
     latent = w.empty_latent_image(extent.initial.width, extent.initial.height, batch)
     model, positive, negative = apply_conditioning(cond, w, comfy, model, clip, style)
@@ -443,7 +443,7 @@ def inpaint(comfy: Client, style: Style, image: Image, mask: Mask, cond: Conditi
     region_expanded = target_bounds.extent.at_least(64).multiple_of(8)
     expanded_bounds = Bounds(*mask.bounds.offset, *region_expanded)
 
-    w = ComfyWorkflow(comfy.nodes_required_inputs)
+    w = ComfyWorkflow(comfy.nodes_inputs)
     model, clip, vae = load_model_with_lora(w, comfy, style)
     in_image = w.load_image(scaled_image)
     in_mask = w.load_mask(scaled_mask)
@@ -521,7 +521,7 @@ def refine(
     extent, image, batch = prepare_image(image, resolve_sd_version(style, comfy), downscale=False)
     sampler_params = _sampler_params(style, live=live, strength=strength)
 
-    w = ComfyWorkflow(comfy.nodes_required_inputs)
+    w = ComfyWorkflow(comfy.nodes_inputs)
     model, clip, vae = load_model_with_lora(w, comfy, style, is_live=live.is_active)
     in_image = w.load_image(image)
     if extent.is_incompatible:
@@ -554,7 +554,7 @@ def refine_region(
     extent, image, mask_image, batch = prepare_masked(image, mask, sd_ver, downscale_if_needed)
     sampler_params = _sampler_params(style, strength=strength, live=live)
 
-    w = ComfyWorkflow(comfy.nodes_required_inputs)
+    w = ComfyWorkflow(comfy.nodes_inputs)
     model, clip, vae = load_model_with_lora(w, comfy, style, is_live=live.is_active)
     in_image = w.load_image(image)
     in_mask = w.load_mask(mask_image)
@@ -647,7 +647,7 @@ def upscale_tiled(
     else:  # SDXL
         tile_extent = Extent(1024, 1024)
 
-    w = ComfyWorkflow(comfy.nodes_required_inputs)
+    w = ComfyWorkflow(comfy.nodes_inputs)
     img = w.load_image(image)
     checkpoint, clip, vae = load_model_with_lora(w, comfy, style)
     upscale_model = w.load_upscale_model(model)

diff --git a/pyproject.toml b/pyproject.toml
@@ -1,3 +1,24 @@
 [tool.black]
 line-length=100
-preview=1
+preview=1
+include='(ai_diffusion|scripts|tests)/.*\.pyi?$'
+extend-exclude='websockets|krita\.pyi$'
+
+[tool.pyright]
+include = [
+    "ai_diffusion",
+    "scripts/*.py",
+    "tests"
+]
+exclude = [
+    "**/__pycache__",
+    "**/.pytest_cache",
+    "**/.server",
+    "ai_diffusion/websockets",
+]
+ignore = [
+    "ai_diffusion/websockets",
+    "krita.pyi"
+]
+extraPaths = ["scripts/typeshed"]
+reportMissingModuleSource = false
diff --git a/requirements.txt b/requirements.txt
@@ -1,5 +1,10 @@
 # This file is for development and running tests.
 # The plugin itself will run inside Krita's embedded Python, and only has access to the Python standard library and Qt5.
+
+# Development
+black
+
+# Testing
 aiohttp
 markdown
 numpy

diff --git a/scripts/docker/Dockerfile b/scripts/docker/Dockerfile
@@ -74,8 +74,8 @@ RUN git clone https://github.com/ltdrdata/ComfyUI-Manager.git /ComfyUI/custom_no
 
 # Copy models
 RUN mkdir -p models
-COPY models/ /models/
-COPY custom_nodes/ComfyUI_IPAdapter_plus/models/* /ComfyUI/custom_nodes/ComfyUI_IPAdapter_plus/models/
+COPY downloads/models/ /models/
+COPY downloads/custom_nodes/ComfyUI_IPAdapter_plus/models/* /ComfyUI/custom_nodes/ComfyUI_IPAdapter_plus/models/
 COPY extra_model_paths.yaml /ComfyUI/
 
 # Install Jupyter