textflowkit 0.1.4__tar.gz → 0.1.5__tar.gz

{textflowkit-0.1.4 → textflowkit-0.1.5}/.gitattributes

@@ -1,22 +1,22 @@
-# Normalise line endings so a build is the same on every platform.
-#
-# Without this, core.autocrlf rewrites text files on Windows checkout, so a
-# wheel built from a Windows clone differs byte-for-byte from one built on
-# Linux - same code, different bytes - and a published hash cannot be
-# reproduced. Text is stored as LF in the repository and in every build.
-* text=auto eol=lf
-
-# Binary assets must never be rewritten.
-*.png binary
-*.jpg binary
-*.jpeg binary
-*.gif binary
-*.ico binary
-*.pdf binary
-*.docx binary
-*.whl binary
-*.gz binary
-
-# Windows-only scripts keep CRLF.
-*.bat text eol=crlf
-*.cmd text eol=crlf
+# Normalise line endings so a build is the same on every platform.
+#
+# Without this, core.autocrlf rewrites text files on Windows checkout, so a
+# wheel built from a Windows clone differs byte-for-byte from one built on
+# Linux - same code, different bytes - and a published hash cannot be
+# reproduced. Text is stored as LF in the repository and in every build.
+* text=auto eol=lf
+
+# Binary assets must never be rewritten.
+*.png binary
+*.jpg binary
+*.jpeg binary
+*.gif binary
+*.ico binary
+*.pdf binary
+*.docx binary
+*.whl binary
+*.gz binary
+
+# Windows-only scripts keep CRLF.
+*.bat text eol=crlf
+*.cmd text eol=crlf

{textflowkit-0.1.4 → textflowkit-0.1.5}/.github/workflows/ci.yml

@@ -57,7 +57,8 @@ jobs:
         shell: bash
         run: |
           python -m pip install --upgrade pip
-          python -m pip install -e ".[dev,mcp,http]"
+          python -m pip install ./packages/textflowkit-fonts
+          python -m pip install -e ".[dev,mcp,http,export]"
 
       - name: Show ffmpeg version
         shell: bash
@@ -84,6 +85,7 @@ jobs:
         run: |
           python -m pip install "setuptools>=83" "build>=1.2"
           python -m build --wheel
+          python -m build --wheel --outdir dist packages/textflowkit-fonts
       - name: Smoke all three installed entry points
         shell: bash
-        run: python scripts/smoke_installed_wheel.py dist/*.whl
+        run: python scripts/smoke_installed_wheel.py dist/textflowkit-*.whl dist/textflowkit_fonts-*.whl

textflowkit-0.1.5/.github/workflows/publish-pypi.yml

@@ -0,0 +1,114 @@
+name: Publish PyPI
+
+on:
+  push:
+    tags: ['v*']
+
+permissions:
+  contents: read
+
+jobs:
+  build:
+    if: github.repository == 'scottconverse/textflowkit'
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
+        with:
+          persist-credentials: false
+          fetch-depth: 0
+      - uses: actions/setup-python@5fda3b95a4ea91299a34e894583c3862153e4b97 # v7.0.0
+        with:
+          python-version: '3.12'
+      - run: python -m pip install packaging
+      - name: Verify release tag and package versions
+        env:
+          RELEASE_TAG: ${{ github.ref_name }}
+        run: |
+          python - <<'PY'
+          import os, tomllib
+          from pathlib import Path
+          from packaging.version import Version
+
+          tag = os.environ['RELEASE_TAG']
+          if not tag.startswith('v'):
+              raise SystemExit('Release tag must start with v')
+          version = str(Version(tag[1:]))
+          if tag != 'v' + version:
+              raise SystemExit('Release tag is not a normalized final version')
+          if Version(version).is_prerelease or Version(version).is_devrelease:
+              raise SystemExit('PyPI release must be a final version')
+          for file in (Path('pyproject.toml'), Path('packages/textflowkit-fonts/pyproject.toml')):
+              declared = tomllib.loads(file.read_text(encoding='utf-8'))['project']['version']
+              if declared != version:
+                  raise SystemExit(f'{file}: {declared} != {version}')
+          PY
+          git merge-base --is-ancestor "$GITHUB_SHA" origin/main || {
+            echo 'Release tag is not merged into main' >&2
+            exit 1
+          }
+      - name: Build distributions
+        run: |
+          python -m pip install --upgrade build packaging
+          python -m build --sdist --wheel --outdir dist/main .
+          python -m build --sdist --wheel --outdir dist/fonts packages/textflowkit-fonts
+          python -m pip install twine
+          python -m twine check dist/main/* dist/fonts/*
+      - uses: actions/upload-artifact@043fb46d1a93c77aae656e7c1c64a875d1fc6a0a # v7.0.1
+        with:
+          name: distributions
+          path: dist/
+          if-no-files-found: error
+
+  publish-fonts:
+    needs: build
+    runs-on: ubuntu-latest
+    environment: pypi
+    permissions:
+      id-token: write
+    steps:
+      - uses: actions/download-artifact@3e5f45b2cfb9172054b4087a40e8e0b5a5461e7c # v8.0.1
+        with:
+          name: distributions
+          path: dist/
+      - uses: pypa/gh-action-pypi-publish@dc37677b2e1c63e2034f94d8a5b11f265b73ba33 # v1.14.2
+        with:
+          packages-dir: dist/fonts/
+
+  publish-main:
+    needs: publish-fonts
+    runs-on: ubuntu-latest
+    environment: pypi
+    permissions:
+      id-token: write
+    steps:
+      - uses: actions/download-artifact@3e5f45b2cfb9172054b4087a40e8e0b5a5461e7c # v8.0.1
+        with:
+          name: distributions
+          path: dist/
+      - uses: pypa/gh-action-pypi-publish@dc37677b2e1c63e2034f94d8a5b11f265b73ba33 # v1.14.2
+        with:
+          packages-dir: dist/main/
+
+  publish-github-release:
+    needs: publish-main
+    runs-on: ubuntu-latest
+    permissions:
+      contents: write
+    steps:
+      - uses: actions/download-artifact@3e5f45b2cfb9172054b4087a40e8e0b5a5461e7c # v8.0.1
+        with:
+          name: distributions
+          path: dist/
+      - name: Create draft with the exact PyPI distributions
+        env:
+          GH_TOKEN: ${{ github.token }}
+          RELEASE_TAG: ${{ github.ref_name }}
+        run: |
+          gh release create "$RELEASE_TAG" dist/main/* dist/fonts/* \
+            --repo "$GITHUB_REPOSITORY" --verify-tag --draft \
+            --title "$RELEASE_TAG" --generate-notes
+      - name: Publish complete GitHub release
+        env:
+          GH_TOKEN: ${{ github.token }}
+          RELEASE_TAG: ${{ github.ref_name }}
+        run: gh release edit "$RELEASE_TAG" --repo "$GITHUB_REPOSITORY" --draft=false

{textflowkit-0.1.4 → textflowkit-0.1.5}/.gitignore

@@ -17,6 +17,7 @@ work/
 out/
 outputs/
 *.wav
+!src/textflowkit/assets/selftest-speech.wav
 *.mp3
 *.m4a
 *.mp4
@@ -34,4 +35,3 @@ Thumbs.db
 .vscode/
 .idea/
 board/
-

{textflowkit-0.1.4 → textflowkit-0.1.5}/.pre-commit-config.yaml

@@ -15,4 +15,4 @@ repos:
       - id: check-yaml
       - id: check-toml
       - id: check-added-large-files
-        args: [--maxkb=1024]
+        args: [--maxkb=1024]

{textflowkit-0.1.4 → textflowkit-0.1.5}/CONTRIBUTING.md

@@ -13,7 +13,8 @@ cd textflowkit
 python -m venv .venv
 # Windows: .venv\Scripts\activate
 # POSIX:   source .venv/bin/activate
-pip install -e ".[dev,mcp,http]"
+pip install ./packages/textflowkit-fonts  # local optional PDF font companion
+pip install -e ".[dev,mcp,http,export]"
 ```
 
 ## Test and lint — the commands CI runs
@@ -60,6 +61,8 @@ mark it so it can be skipped in CI.
 - Add a test for a bug fix. If you cannot, say why in the PR description.
 - Run `ruff check .` and `python -m pytest` before opening the PR.
 - Do not commit media files, transcripts, or scratch output. `work/` is gitignored.
+  The bundled `src/textflowkit/assets/selftest-speech.wav` is an intentional,
+  small synthetic speech fixture; do not replace it with private recordings.
 
 ## Evidence language
 

{textflowkit-0.1.4 → textflowkit-0.1.5}/LEGAL.md

@@ -41,4 +41,4 @@ outside the intended use of this project.
 
 "textflowkit" is an independent open-source project. It is not affiliated with,
 endorsed by, or connected to any other product or organisation using a similar
-name.
+name.

{textflowkit-0.1.4 → textflowkit-0.1.5}/LICENSE

@@ -200,4 +200,4 @@
    WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
    See the License for the specific language governing permissions and
    limitations under the License.
-
+

{textflowkit-0.1.4 → textflowkit-0.1.5}/PKG-INFO

@@ -1,9 +1,12 @@
 Metadata-Version: 2.5
 Name: textflowkit
-Version: 0.1.4
+Version: 0.1.5
 Summary: Cross-platform media transcription toolkit: URL or file in, timestamped transcripts and subtitle files out.
-Project-URL: Homepage, https://github.com/scottconverse/textflowkit
+Project-URL: Homepage, https://www.textflowkit.org/
 Project-URL: Issues, https://github.com/scottconverse/textflowkit/issues
+Project-URL: Source, https://github.com/scottconverse/textflowkit
+Project-URL: Documentation, https://github.com/scottconverse/textflowkit/tree/main/docs
+Project-URL: Changelog, https://github.com/scottconverse/textflowkit/releases
 License-Expression: Apache-2.0
 License-File: LICENSE
 Keywords: speech-to-text,srt,subtitles,transcription,vtt,whisper,yt-dlp
@@ -27,6 +30,7 @@ Requires-Dist: pyannote-audio>=4.0; extra == 'all'
 Requires-Dist: pydantic>=2.7; extra == 'all'
 Requires-Dist: python-docx>=1.1; extra == 'all'
 Requires-Dist: reportlab>=4.0; extra == 'all'
+Requires-Dist: textflowkit-fonts<0.2,>=0.1.5; extra == 'all'
 Requires-Dist: uvicorn>=0.30; extra == 'all'
 Provides-Extra: dev
 Requires-Dist: pypdf>=5; extra == 'dev'
@@ -38,6 +42,7 @@ Requires-Dist: pyannote-audio>=4.0; extra == 'diarize'
 Provides-Extra: export
 Requires-Dist: python-docx>=1.1; extra == 'export'
 Requires-Dist: reportlab>=4.0; extra == 'export'
+Requires-Dist: textflowkit-fonts<0.2,>=0.1.5; extra == 'export'
 Provides-Extra: http
 Requires-Dist: fastapi>=0.115; extra == 'http'
 Requires-Dist: pydantic>=2.7; extra == 'http'
@@ -139,6 +144,47 @@ python -m pip install 'textflowkit[mcp]'    # MCP server
 python -m pip install 'textflowkit[http]'   # JSON HTTP API
 ```
 
+PDF/DOCX export is optional. The default wheel stays small; the `export` extra
+installs `textflowkit-fonts` for offline multilingual PDF rendering:
+
+```bash
+python -m pip install 'textflowkit[export]'
+```
+
+## Python API
+
+The same pipeline used by the CLI and adapters is available to Python callers:
+
+```python
+from textflowkit import transcribe
+
+result = transcribe(
+    "meeting.mp4",            # also accepts supported URLs
+    model="small",
+    formats=["json", "srt", "txt"],
+    output_dir="transcripts", # omit to return the transcript without writing files
+)
+print(result.transcript.text)
+print(result.transcript.duration)  # full media duration in seconds
+print(result.outputs)              # pathlib.Path objects for written files
+for segment in result.transcript.segments:
+    print(segment.start, segment.end, segment.speaker, segment.text)
+    for word in segment.words:
+        print("  ", word.start, word.end, word.text)
+```
+
+`transcribe()` returns `TranscribeResult` with a canonical `Transcript` and
+written output paths. `Transcript.to_dict()` / `.to_json()` preserve segment and
+word timing; older transcript JSON without `words` remains readable. Pass
+`input_root=` to confine local input paths for untrusted callers. See
+[the install guide](https://github.com/scottconverse/textflowkit/blob/main/docs/install.md)
+for ffmpeg and Windows ROCm setup.
+
+MCP and HTTP transcript reads omit word timings by default to keep responses
+small; set `include_words=true` on a JSON read to receive them. Saved files,
+Python results, and durable job records still retain the source-language words,
+including when segment text has been translated.
+
 **AMD ROCm on native Windows:** do not use the generic command in an environment
 with a working ROCm PyTorch install. Ordinary dependency resolution can replace
 that torch build. Follow the [ROCm install notes](https://github.com/scottconverse/textflowkit/blob/main/docs/install.md)

{textflowkit-0.1.4 → textflowkit-0.1.5}/README.md

@@ -91,6 +91,47 @@ python -m pip install 'textflowkit[mcp]'    # MCP server
 python -m pip install 'textflowkit[http]'   # JSON HTTP API
 ```
 
+PDF/DOCX export is optional. The default wheel stays small; the `export` extra
+installs `textflowkit-fonts` for offline multilingual PDF rendering:
+
+```bash
+python -m pip install 'textflowkit[export]'
+```
+
+## Python API
+
+The same pipeline used by the CLI and adapters is available to Python callers:
+
+```python
+from textflowkit import transcribe
+
+result = transcribe(
+    "meeting.mp4",            # also accepts supported URLs
+    model="small",
+    formats=["json", "srt", "txt"],
+    output_dir="transcripts", # omit to return the transcript without writing files
+)
+print(result.transcript.text)
+print(result.transcript.duration)  # full media duration in seconds
+print(result.outputs)              # pathlib.Path objects for written files
+for segment in result.transcript.segments:
+    print(segment.start, segment.end, segment.speaker, segment.text)
+    for word in segment.words:
+        print("  ", word.start, word.end, word.text)
+```
+
+`transcribe()` returns `TranscribeResult` with a canonical `Transcript` and
+written output paths. `Transcript.to_dict()` / `.to_json()` preserve segment and
+word timing; older transcript JSON without `words` remains readable. Pass
+`input_root=` to confine local input paths for untrusted callers. See
+[the install guide](https://github.com/scottconverse/textflowkit/blob/main/docs/install.md)
+for ffmpeg and Windows ROCm setup.
+
+MCP and HTTP transcript reads omit word timings by default to keep responses
+small; set `include_words=true` on a JSON read to receive them. Saved files,
+Python results, and durable job records still retain the source-language words,
+including when segment text has been translated.
+
 **AMD ROCm on native Windows:** do not use the generic command in an environment
 with a working ROCm PyTorch install. Ordinary dependency resolution can replace
 that torch build. Follow the [ROCm install notes](https://github.com/scottconverse/textflowkit/blob/main/docs/install.md)

{textflowkit-0.1.4 → textflowkit-0.1.5}/docs/adapters.md

@@ -136,7 +136,7 @@ textflowkit-http --host 127.0.0.1 --port 8767
 | POST | `/jobs/{id}/resume` | resume a saved durable request/checkpoint |
 | GET | `/jobs` | list recent jobs |
 | GET | `/jobs/{id}` | job status |
-| GET | `/jobs/{id}/transcript?format=&offset=&limit=&start=&end=` | rendered transcript, optionally sliced |
+| GET | `/jobs/{id}/transcript?format=&offset=&limit=&start=&end=&include_words=` | rendered transcript, optionally sliced; word timings are opt-in for JSON |
 | GET | `/jobs/{id}/search?q=&limit=&context=` | search a transcript |
 | POST | `/jobs/{id}/export?formats=docx&formats=pdf` | write files to disk (docx/pdf included) |
 | POST | `/jobs/{id}/cancel` | request cancellation |
@@ -338,6 +338,15 @@ question being asked.
 | `offset` | skip this many segments **within the selected range** |
 | `limit` | return at most this many |
 | `start` / `end` | restrict by time in seconds (inclusive) |
+| `include_words` | include source-language word timings in JSON; default `false` on MCP and HTTP reads |
+
+Saved JSON, SQLite job records, and resume checkpoints retain word timings even
+when `include_words=false`; the option reduces response size, not storage size.
+In one short reviewer sample, transcript JSON grew from 508 to 1,641 bytes
+(roughly threefold); the multiplier varies with segment and word counts.
+If the transcript was translated, the optional word timings still refer to the
+**original spoken language**, not the translated segment text. The Python API
+also retains the original word timings.
 
 Filtering is time first, then offset/limit inside that window - `offset` counts
 from the start of the requested range, not the start of the transcript. The

{textflowkit-0.1.4 → textflowkit-0.1.5}/docs/install.md

@@ -133,7 +133,9 @@ textflowkit selftest --skip-transcribe   # compute device only, no model downloa
 ```
 
 It runs a real matmul on the selected device and then a real Whisper pass over a
-generated probe clip, printing PASS/FAIL for each and naming the torch build. A
+bundled synthetic speech clip. It fails if Whisper returns no nonempty, timed
+speech segment. A PASS proves text generation, not transcription accuracy.
+It prints PASS/FAIL for each stage and names the torch build. A
 ROCm install reports `torch <ver>+rocm*` and the device name; a stock CPU wheel
 reports plain `torch <ver>`.
 
@@ -235,4 +237,3 @@ how many segments were labelled. If the backend or token is missing, the run
 
 With no GPU, the engine selects CPU automatically. Pass `--device cpu` to force it.
 CPU transcription is dramatically slower; prefer a smaller `--model`.
-

{textflowkit-0.1.4 → textflowkit-0.1.5}/docs/release-checklist.md

@@ -51,13 +51,36 @@ receipt. It does not upload cookies, media, or transcripts to GitHub.
 
 ## PyPI publication
 
-If this release is also published to PyPI, upload the **same** verified wheel
-and source archive as the GitHub release. Check package metadata before upload,
-then compare both PyPI SHA-256 digests with the GitHub release assets. Finally,
-install that exact version from the public PyPI index in a clean environment
-and smoke the CLI entry point. An install with `--no-deps` proves distribution
-and entry-point wiring only; it does not prove a full transcription run.
+After merged-main CI and the local release checks pass, push an annotated final
+version tag (for example `v0.1.5`) on that verified `main` commit. Do **not**
+create or publish a GitHub release manually. The tag push triggers
+[`publish-pypi.yml`](../.github/workflows/publish-pypi.yml). It verifies the
+tag, both package versions, and ancestry on `main`, builds the four artifacts,
+publishes `textflowkit-fonts` first and `textflowkit` second using PyPI Trusted
+Publishing, then creates a **draft** GitHub release with those exact artifacts.
+Only after the assets are attached does it make that release public.
+No upload token is passed to CI. Both PyPI projects must have trusted publishers
+for owner `scottconverse`, repository `textflowkit`, workflow
+`publish-pypi.yml`, and environment `pypi`. The GitHub `pypi` environment
+should require maintainer approval before an upload job can proceed.
+
+This is not a transaction across two PyPI projects and GitHub. If fonts upload
+succeeds but core fails, **there is no public GitHub release**, but fonts are
+already on PyPI. Check whether any core files reached PyPI before retrying:
+PyPI will reject duplicate filenames, and this workflow intentionally does not
+silently skip them. Correct the cause and rerun only failed jobs when safe; if
+some core files are present, reconcile their hashes against the original build
+artifact and finish the missing files deliberately. If both PyPI uploads succeed
+but draft creation or publishing fails, do not republish either PyPI project;
+complete the GitHub draft and assets from the exact build artifact after hash
+verification. Do not move the published version tag or claim a complete release
+until all four PyPI files and all four GitHub assets match.
+
+After publication, compare PyPI SHA-256 digests against the GitHub release assets
+for **both** packages, and clean-install the exact version with
+`textflowkit[export,mcp,http]`. Run `doctor`, `selftest`, and an actual PDF export.
+An install with `--no-deps` only proves distribution wiring, not transcription.
 
 Never place an API token in this repository, a CI log, or a shell command line.
 An upload is a separate public release action; passing CI alone does not
-authorize it.
+authorize it. Do not call a release complete if either package upload failed.

{textflowkit-0.1.4 → textflowkit-0.1.5}/docs/roadmap.md

@@ -73,6 +73,17 @@ transcription service.
 - [x] Prepare v0.1.4 with a PyPI-first package description and install guide
 - [x] Make PyPI the default developer install path while keeping GitHub
   releases and the native-Windows ROCm dependency instructions
+- [ ] Ship v0.1.5 review follow-ups: speech-bearing self-test, full media
+  duration, retained word timings, Python API documentation, richer PyPI
+  project links, a small core wheel with optional offline fonts, and a
+  tokenless Trusted Publishing release workflow for both packages
+- [ ] Follow-up: normalize exported file permissions on POSIX to respect the
+  process umask. Current atomic temporary files can leave outputs mode `0600`,
+  preventing another account (such as a separate web server user) from reading
+  exported subtitles or documents. Fix and regression-test separately.
+- [ ] Follow-up: decouple the fonts package's version from core releases so
+  unchanged font wheels are not rebuilt/uploaded every patch. Keep duplicate
+  PyPI uploads visible rather than relying on `skip-existing`.
 
 The 13 listed media platforms are recognised through `yt-dlp`; **only YouTube**
 has an opt-in [live URL transcription release gate](release-checklist.md), not

{textflowkit-0.1.4 → textflowkit-0.1.5}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "hatchling.build"
 
 [project]
 name = "textflowkit"
-version = "0.1.4"
+version = "0.1.5"
 description = "Cross-platform media transcription toolkit: URL or file in, timestamped transcripts and subtitle files out."
 readme = "README.md"
 requires-python = ">=3.10"
@@ -36,7 +36,7 @@ http = ["fastapi>=0.115", "uvicorn>=0.30", "pydantic>=2.7"]
 # no longer provides. See docs/install.md before installing - pyannote pulls
 # torch open-endedly, so pin your torch stack or you lose GPU acceleration.
 diarize = ["pyannote.audio>=4.0"]
-export = ["python-docx>=1.1", "reportlab>=4.0"]
+export = ["python-docx>=1.1", "reportlab>=4.0", "textflowkit-fonts>=0.1.5,<0.2"]
 all = ["textflowkit[mcp,http,diarize,export]"]
 
 [project.scripts]
@@ -45,11 +45,18 @@ textflowkit-mcp = "textflowkit.adapters.mcp_server:main"
 textflowkit-http = "textflowkit.adapters.http_server:main"
 
 [project.urls]
-Homepage = "https://github.com/scottconverse/textflowkit"
+Homepage = "https://www.textflowkit.org/"
 Issues = "https://github.com/scottconverse/textflowkit/issues"
+Source = "https://github.com/scottconverse/textflowkit"
+Documentation = "https://github.com/scottconverse/textflowkit/tree/main/docs"
+Changelog = "https://github.com/scottconverse/textflowkit/releases"
 
 [tool.hatch.build.targets.wheel]
 packages = ["src/textflowkit"]
+artifacts = ["src/textflowkit/assets/*.wav"]
+
+[tool.hatch.build.targets.sdist]
+exclude = ["/packages/textflowkit-fonts"]
 
 [tool.ruff]
 line-length = 100

{textflowkit-0.1.4 → textflowkit-0.1.5}/scripts/smoke_installed_wheel.py

@@ -19,6 +19,7 @@ import time
 import urllib.error
 import urllib.request
 import venv
+import zipfile
 from pathlib import Path
 
 
@@ -151,9 +152,19 @@ def http_smoke(command: Path, cwd: Path, env: dict[str, str], *, production: boo
 
 
 def main() -> None:
-    if len(sys.argv) != 2:
-        raise SystemExit("usage: python scripts/smoke_installed_wheel.py PATH_TO_WHEEL")
+    if len(sys.argv) != 3:
+        raise SystemExit("usage: python scripts/smoke_installed_wheel.py MAIN_WHEEL FONTS_WHEEL")
     wheel = Path(sys.argv[1]).resolve(strict=True)
+    font_wheel = Path(sys.argv[2]).resolve(strict=True)
+    with zipfile.ZipFile(wheel) as archive:
+        assert not any(name.endswith(".ttf") for name in archive.namelist())
+        assert "textflowkit/assets/selftest-speech.wav" in archive.namelist()
+    with zipfile.ZipFile(font_wheel) as archive:
+        assert all(
+            f"textflowkit_fonts/fonts/{name}.ttf" in archive.namelist()
+            for name in ("NotoSans", "NotoSansArabic", "NotoSansSC")
+        )
+        assert any("OFL-NotoSans.txt" in name for name in archive.namelist())
     with tempfile.TemporaryDirectory(prefix="tfk-wheel-smoke-") as directory:
         root = Path(directory)
         env_root = root / "venv"
@@ -165,13 +176,33 @@ def main() -> None:
         env.pop("TEXTFLOWKIT_DB", None)
         env["TEXTFLOWKIT_OUTPUT_ROOT"] = str(root / "outputs")
         run(py, "-m", "pip", "install", "--disable-pip-version-check", "--no-deps",
-            wheel, cwd=root, env=env)
+            wheel, font_wheel, cwd=root, env=env)
         run(py, "-m", "pip", "install", "--disable-pip-version-check",
             "mcp>=2.0", "fastapi>=0.115", "uvicorn>=0.30", "pydantic>=2.7",
+            "reportlab>=4.0", "pypdf>=5",
             cwd=root, env=env)
         installed_path = run(py, "-c", "import textflowkit; print(textflowkit.__file__)",
                              cwd=root, env=env).strip()
         assert str(root).casefold() in installed_path.casefold(), installed_path
+        speech_fixture = run(
+            py, "-c",
+            "from importlib.resources import files; "
+            "p=files('textflowkit').joinpath('assets/selftest-speech.wav'); "
+            "assert p.read_bytes().startswith(b'RIFF'); print('speech fixture present')",
+            cwd=root, env=env,
+        )
+        assert "speech fixture present" in speech_fixture
+        pdf_smoke = run(
+            py, "-c",
+            "from textflowkit import Segment, Transcript; "
+            "from textflowkit.render import render_bytes; "
+            "from pypdf import PdfReader; import io; "
+            "blob=render_bytes(Transcript(source='wheel',segments=[Segment(0,1,'Hello 你好')]),'pdf'); "
+            "assert '你好' in PdfReader(io.BytesIO(blob)).pages[0].extract_text(); "
+            "print('PDF export passed')",
+            cwd=root, env=env,
+        )
+        assert "PDF export passed" in pdf_smoke
         cli = executable(env_root, "textflowkit")
         mcp = executable(env_root, "textflowkit-mcp")
         http = executable(env_root, "textflowkit-http")

textflowkit-0.1.5/src/textflowkit/__init__.py

@@ -0,0 +1,8 @@
+"""textflowkit - cross-platform media transcription toolkit."""
+
+from textflowkit.core.model import Segment, Transcript, WordTiming
+from textflowkit.core.pipeline import PipelineError, transcribe
+
+__version__ = "0.1.5"
+
+__all__ = ["PipelineError", "Segment", "Transcript", "WordTiming", "__version__", "transcribe"]

{textflowkit-0.1.4 → textflowkit-0.1.5}/src/textflowkit/adapters/__init__.py

@@ -8,4 +8,4 @@ here, so the MCP server, the HTTP API, and any future frontend cannot drift.
 - `http_server` - JSON HTTP API for software products and web frontends
 """
 
-__all__ = ["http_server", "mcp_server"]
+__all__ = ["http_server", "mcp_server"]

{textflowkit-0.1.4 → textflowkit-0.1.5}/src/textflowkit/adapters/http_server.py

@@ -283,11 +283,14 @@ def get_transcript(
     limit: int | None = None,
     start: float | None = None,
     end: float | None = None,
+    include_words: bool = False,
 ):
     """Transcript for a completed job, optionally a slice.
 
     `offset`/`limit` page through segments; `start`/`end` select a time range in
-    seconds. The JSON form reports total_segments and has_more.
+    seconds. The JSON form reports total_segments and has_more. Word timings
+    are omitted unless include_words is true; translated segments retain
+    source-language word timings.
     """
     job, tr = _finished_transcript(job_id)
     if production_enabled():
@@ -325,7 +328,7 @@ def get_transcript(
         response = {
             "job_id": job.id,
             **page.as_dict(),
-            "transcript": sliced.to_dict(),
+            "transcript": sliced.to_dict(include_words=include_words),
         }
         try:
             enforce_output_limit(len(json.dumps(response, ensure_ascii=False).encode("utf-8")))
@@ -461,5 +464,3 @@ def main(argv: list[str] | None = None) -> int:
 if __name__ == "__main__":
     raise SystemExit(main())
 
-
-

{textflowkit-0.1.4 → textflowkit-0.1.5}/src/textflowkit/adapters/mcp_server.py

@@ -248,6 +248,7 @@ def get_transcript(
     limit: int | None = None,
     start: float | None = None,
     end: float | None = None,
+    include_words: bool = False,
 ) -> dict[str, Any]:
     """Read the transcript for a completed job, optionally a slice of it.
 
@@ -262,6 +263,9 @@ def get_transcript(
         limit: Return at most this many segments.
         start: Only segments ending at or after this time (seconds).
         end: Only segments starting at or before this time (seconds).
+        include_words: Include source-language word timings in JSON output.
+            Defaults to false to keep model context small; translated text
+            does not produce translated word timings.
     """
     job, err = _resolve_job(job_id)
     if err:
@@ -307,7 +311,8 @@ def get_transcript(
         "language": tr.language,
         "platform": tr.platform,
         **page.as_dict(),
-        "content": render(sliced, norm),
+        "content": sliced.to_json(include_words=include_words) if norm == "json"
+        else render(sliced, norm),
     }
     if page.has_more:
         payload["next"] = (
@@ -551,4 +556,3 @@ if __name__ == "__main__":
 
 
 
-