initial commit

.gitignore

@@ -0,0 +1,3 @@
+shell.nix
+__pycache__/
+venv/

.mypy.ini

@@ -0,0 +1,12 @@
+[mypy]
+files = .
+exclude = build
+namespace_packages = True
+show_error_codes = True
+strict = True
+enable_error_code = redundant-expr, truthy-bool, ignore-without-code
+; plugins = numpy.typing.mypy_plugin
+ignore_missing_imports = True
+no_implicit_reexport = False
+disallow_untyped_calls = False
+ignore_missing_imports = True

README.md

@@ -0,0 +1,10 @@
+---
+title: OCR
+emoji: 🔥
+colorFrom: green
+colorTo: blue
+sdk: gradio
+sdk_version: 4.36.1
+app_file: app.py
+pinned: false
+---

app.py

@@ -0,0 +1,50 @@
+from __future__ import annotations
+
+import os
+from pathlib import Path
+
+import gradio as gr  # type: ignore
+import rerun as rr
+import rerun.blueprint as rrb
+from gradio_rerun import Rerun  # type: ignore
+from ocr import detect_and_log_layout  # type: ignore
+
+
+@rr.thread_local_stream("PaddleOCR")
+def log_to_rr(img_path: Path):
+    print(img_path)
+    stream = rr.binary_stream()
+
+    blueprint = rrb.Blueprint(
+        rrb.Vertical(
+            rrb.Spatial2DView(name="Input", contents=["Image/**"]),
+        ),
+        collapse_panels=True,
+    )
+    rr.send_blueprint(blueprint)
+
+    detect_and_log_layout(img_path)
+
+    yield stream.read()
+
+
+with gr.Blocks() as demo:
+    with gr.Row():
+        with gr.Column(scale=1):
+            with gr.Row():
+                input_image = gr.Image(label="Input Image", image_mode="RGBA", sources="upload", type="filepath")
+            with gr.Row():
+                button = gr.Button()
+            with gr.Row():
+                gr.Examples(
+                    examples=[os.path.join("examples", img_name) for img_name in sorted(os.listdir("examples"))],
+                    inputs=[input_image],
+                    label="Examples",
+                    cache_examples=False,
+                    examples_per_page=12,
+                )
+        with gr.Column(scale=4):
+            viewer = Rerun(streaming=True, height=900)
+    button.click(log_to_rr, inputs=[input_image], outputs=[viewer])
+
+demo.launch()

lychee.toml

@@ -0,0 +1,82 @@
+# Copied from https://github.com/rerun-io/rerun_template
+
+################################################################################
+# Config for the link checker lychee.
+#
+# Download & learn more at:
+# https://github.com/lycheeverse/lychee
+#
+# Example config:
+# https://github.com/lycheeverse/lychee/blob/master/lychee.example.toml
+#
+# Run `lychee . --dump` to list all found links that are being checked.
+#
+# Note that by default lychee will only check markdown and html files,
+# to check any other files you have to point to them explicitly, e.g.:
+# `lychee **/*.rs`
+# To make things worse, `exclude_path` is ignored for these globs,
+# so local runs with lots of gitignored files will be slow.
+# (https://github.com/lycheeverse/lychee/issues/1405)
+#
+# This unfortunately doesn't list anything for non-glob checks.
+################################################################################
+
+# Maximum number of concurrent link checks.
+# Workaround for "too many open files" error on MacOS, see https://github.com/lycheeverse/lychee/issues/1248
+max_concurrency = 32
+
+# Check links inside `<code>` and `<pre>` blocks as well as Markdown code blocks.
+include_verbatim = true
+
+# Proceed for server connections considered insecure (invalid TLS).
+insecure = true
+
+# Exclude these filesystem paths from getting checked.
+exclude_path = [
+  # Unfortunately lychee doesn't yet read .gitignore https://github.com/lycheeverse/lychee/issues/1331
+  # The following entries are there because of that:
+  ".git",
+  "__pycache__",
+  "_deps/",
+  ".pixi",
+  "build",
+  "target_ra",
+  "target_wasm",
+  "target",
+  "venv",
+]
+
+# Exclude URLs and mail addresses from checking (supports regex).
+exclude = [
+  # Skip speculative links
+  '.*?speculative-link',
+
+  # Strings with replacements.
+  '/__VIEWER_VERSION__/', # Replacement variable __VIEWER_VERSION__.
+  '/\$',                  # Replacement variable $.
+  '/GIT_HASH/',           # Replacement variable GIT_HASH.
+  '\{\}',                 # Ignore links with string interpolation.
+  '\$relpath\^',          # Relative paths as used by rerun_cpp's doc header.
+  '%7B.+%7D',             # Ignore strings that look like ready to use links but contain a replacement strings. The URL escaping is for '{.+}' (this seems to be needed for html embedded urls since lychee assumes they use this encoding).
+  '%7B%7D',               # Ignore links with string interpolation, escaped variant.
+
+  # Local links that require further setup.
+  'http://127.0.0.1',
+  'http://localhost',
+  'recording:/',      # rrd recording link.
+  'ws:/',
+  're_viewer.js',     # Build artifact that html is linking to.
+
+  # Api endpoints.
+  'https://fonts.googleapis.com/', # Font API entrypoint, not a link.
+  'https://fonts.gstatic.com/',    # Font API entrypoint, not a link.
+  'https://tel.rerun.io/',         # Analytics endpoint.
+
+  # Avoid rate limiting.
+  'https://crates.io/crates/.*',                  # Avoid crates.io rate-limiting
+  'https://github.com/rerun-io/rerun/commit/\.*', # Ignore links to our own commits (typically in changelog).
+  'https://github.com/rerun-io/rerun/pull/\.*',   # Ignore links to our own pull requests (typically in changelog).
+
+  # Used in rerun_template repo until the user search-replaces `new_repo_name`
+  'https://github.com/rerun-io/new_repo_name',
+]

pyproject.toml

@@ -0,0 +1,72 @@
+# Copied from https://github.com/rerun-io/rerun_template
+
+[tool.ruff]
+# https://beta.ruff.rs/docs/configuration/
+
+target-version = "py38"
+
+# Enable unsafe fixes to allow ruff to apply fixes that may change the behavior of the code.
+# This is needed because otherwise ruff will not be able to trim whitespaces in (codegened) docstrings.
+unsafe-fixes = true
+
+# Allow preview lints to be enabled (like `PLW1514` to force `encoding` on open).
+preview = true
+# But we only want to opt-in to certain preview rules!
+lint.explicit-preview-rules = true
+
+extend-exclude = [
+    # Automatically generated test artifacts
+    "venv/",
+    "target/",
+]
+
+lint.ignore = [
+    # These makes sense to ignore in example code, but for a proper library we should not ignore these.
+    "D100", # Missing docstring in public module
+    "D101", # Missing docstring in public class
+    "D103", # Missing docstring in public function
+
+    # No blank lines allowed after function docstring.
+    "D202",
+
+    # npydocstyle: http://www.pydocstyle.org/en/stable/error_codes.html
+    # numpy convention with a few additional lints
+    "D107",
+    "D203",
+    "D212",
+    "D401",
+    "D402",
+    "D415",
+    "D416",
+
+    # Ruff can't fix this error on its own (yet)
+    # Having ruff check this causes errors that prevent the code-formatting process from completing.
+    "E501",
+
+    # allow relative imports
+    "TID252",
+
+    "UP007", # We need this, or `ruff format` will convert `Union[X, Y]` to `X | Y` which break on Python 3.8
+]
+
+line-length = 120
+lint.select = [
+    "D",       # pydocstyle codes https://www.pydocstyle.org/en/latest/error_codes.html
+    "E",       # pycodestyle error codes: https://pycodestyle.pycqa.org/en/latest/intro.html#error-codes
+    "F",       # Flake8 error codes https://flake8.pycqa.org/en/latest/user/error-codes.html
+    "I",       # Isort
+    "TID",     # flake8-tidy-imports
+    "W",       # pycodestyle warning codes: https://pycodestyle.pycqa.org/en/latest/intro.html#error-codes
+    "UP",      # pyupgrade (ensures idomatic code for supported python version)
+    "PLW1514", # Force setting `encoding` for open calls. This is in order to prevent issues when opening utf8 files on windows where the default encoding may depend on the active locale. https://docs.astral.sh/ruff/rules/unspecified-encoding/
+]
+
+lint.unfixable = [
+    "PLW1514", # Automatic fix for `encoding` doesn't do what we want - it queries the locale for the preferred encoding which is exactly what we want to avoid.
+]
+
+[tool.ruff.lint.per-file-ignores]
+"__init__.py" = ["F401", "F403"]
+
+[tool.ruff.lint.isort]
+required-imports = ["from __future__ import annotations"]

requirements.txt

@@ -0,0 +1,5 @@
+rerun-sdk>=0.16.0,<0.17.0
+spaces
+gradio
+gradio_rerun
+git+https://github.com/rerun-io/rerun.git@c5f817686c6a1b8723ddeee25a9aed0ff4fa1e92#egg=ocr&subdirectory=examples/python/ocr