ancient-processor 4.0.1__cp314-cp314-win_amd64.whl
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- ancient_processor-4.0.1.dist-info/METADATA +285 -0
- ancient_processor-4.0.1.dist-info/RECORD +47 -0
- ancient_processor-4.0.1.dist-info/WHEEL +5 -0
- ancient_processor-4.0.1.dist-info/entry_points.txt +2 -0
- ancient_processor-4.0.1.dist-info/licenses/LICENSE +21 -0
- ancient_processor-4.0.1.dist-info/top_level.txt +1 -0
- super_processor/__init__.py +53 -0
- super_processor/__main__.py +5 -0
- super_processor/_core.cp314-win_amd64.pyd +0 -0
- super_processor/_core.pyi +23 -0
- super_processor/_version.py +24 -0
- super_processor/cli.py +1145 -0
- super_processor/diagnose.py +193 -0
- super_processor/doctor.py +296 -0
- super_processor/encoders.py +101 -0
- super_processor/estimators.py +570 -0
- super_processor/file_dialog.py +111 -0
- super_processor/gemini.py +1684 -0
- super_processor/gpu_heat.py +147 -0
- super_processor/jobs.py +268 -0
- super_processor/look.py +132 -0
- super_processor/models.py +278 -0
- super_processor/plan.py +143 -0
- super_processor/plans.py +366 -0
- super_processor/preview.py +233 -0
- super_processor/probe.py +459 -0
- super_processor/progress.py +108 -0
- super_processor/py.typed +1 -0
- super_processor/qa.py +150 -0
- super_processor/recipe.py +313 -0
- super_processor/reframe.py +345 -0
- super_processor/regions.py +173 -0
- super_processor/render.py +154 -0
- super_processor/review.py +901 -0
- super_processor/segment_loop.py +494 -0
- super_processor/segments.py +895 -0
- super_processor/shot_plan.py +816 -0
- super_processor/shots.py +578 -0
- super_processor/span_sr.py +429 -0
- super_processor/stability.py +286 -0
- super_processor/templates.py +396 -0
- super_processor/treatments.py +342 -0
- super_processor/upscale.py +2434 -0
- super_processor/validator.py +355 -0
- super_processor/wizard.py +1972 -0
- super_processor/wizard_pages.py +1192 -0
- super_processor/worker.py +325 -0
|
@@ -0,0 +1,285 @@
|
|
|
1
|
+
Metadata-Version: 2.4
|
|
2
|
+
Name: ancient-processor
|
|
3
|
+
Version: 4.0.1
|
|
4
|
+
Summary: Local-first AI-assisted video processing
|
|
5
|
+
Author: Super Processor contributors
|
|
6
|
+
License-Expression: MIT
|
|
7
|
+
Project-URL: Documentation, https://github.com/LKPJohn2026/super-processor/tree/main/docs
|
|
8
|
+
Project-URL: Issues, https://github.com/LKPJohn2026/super-processor/issues
|
|
9
|
+
Project-URL: Repository, https://github.com/LKPJohn2026/super-processor
|
|
10
|
+
Keywords: cython,ffmpeg,video-processing,computer-vision
|
|
11
|
+
Classifier: Development Status :: 3 - Alpha
|
|
12
|
+
Classifier: Environment :: Console
|
|
13
|
+
Classifier: Intended Audience :: Developers
|
|
14
|
+
Classifier: Programming Language :: Cython
|
|
15
|
+
Classifier: Programming Language :: Python :: 3
|
|
16
|
+
Classifier: Programming Language :: Python :: 3 :: Only
|
|
17
|
+
Classifier: Programming Language :: Python :: 3.10
|
|
18
|
+
Classifier: Programming Language :: Python :: 3.11
|
|
19
|
+
Classifier: Programming Language :: Python :: 3.12
|
|
20
|
+
Classifier: Programming Language :: Python :: 3.13
|
|
21
|
+
Classifier: Programming Language :: Python :: 3.14
|
|
22
|
+
Classifier: Programming Language :: Python :: Implementation :: CPython
|
|
23
|
+
Classifier: Typing :: Typed
|
|
24
|
+
Requires-Python: >=3.10
|
|
25
|
+
Description-Content-Type: text/markdown
|
|
26
|
+
License-File: LICENSE
|
|
27
|
+
Provides-Extra: gpu
|
|
28
|
+
Requires-Dist: spandrel<0.5,>=0.4.2; extra == "gpu"
|
|
29
|
+
Requires-Dist: torch>=2.4; extra == "gpu"
|
|
30
|
+
Provides-Extra: dev
|
|
31
|
+
Requires-Dist: Cython<4,>=3.0; extra == "dev"
|
|
32
|
+
Requires-Dist: build>=1.2.2; extra == "dev"
|
|
33
|
+
Requires-Dist: mypy>=1.11; extra == "dev"
|
|
34
|
+
Requires-Dist: pytest>=8.3; extra == "dev"
|
|
35
|
+
Requires-Dist: pytest-cov>=6; extra == "dev"
|
|
36
|
+
Requires-Dist: ruff>=0.11; extra == "dev"
|
|
37
|
+
Requires-Dist: setuptools>=77; extra == "dev"
|
|
38
|
+
Requires-Dist: twine>=5.1; extra == "dev"
|
|
39
|
+
Dynamic: license-file
|
|
40
|
+
|
|
41
|
+
# Super Processor
|
|
42
|
+
|
|
43
|
+
[](https://github.com/LKPJohn2026/super-processor/actions/workflows/ci.yml)
|
|
44
|
+
|
|
45
|
+
Local-first AI video restoration: SPAN 2x upscaling, with a strength
|
|
46
|
+
control for how much new detail it adds.
|
|
47
|
+
|
|
48
|
+
Super Processor enhances existing footage. A localhost wizard has Gemini watch
|
|
49
|
+
a small, silent copy of the video, split it into segments, and choose the
|
|
50
|
+
output size, then plans each segment in turn with bounded settings and a short
|
|
51
|
+
before/after preview you approve or send back with a note. It runs a local
|
|
52
|
+
SPAN restoration upscale (or a plain resize where the upscaler would only
|
|
53
|
+
harm), then lets you revise a time range with a note. Gemini structured output
|
|
54
|
+
only labels, picks numbers inside fixed bounds, and checks previews. FFmpeg
|
|
55
|
+
detects, measures, trims, splices, and encodes from engineering-owned argument
|
|
56
|
+
lists. Models never write shell commands.
|
|
57
|
+
|
|
58
|
+
## Requirements
|
|
59
|
+
|
|
60
|
+
- Python 3.10 or newer and a C compiler (Cython extension)
|
|
61
|
+
- FFmpeg 6 or newer and `ffprobe`, with `libx264`. Every output is H.264:
|
|
62
|
+
restoration delivery uses `h264_nvenc` when a test encode works on this
|
|
63
|
+
machine and `libx264` otherwise, and the CLI export is `libx264` (or an
|
|
64
|
+
H.264 hardware encoder named with `--encoder`). `super-processor doctor`
|
|
65
|
+
requires `libx264` and reports whether NVENC and CUDA decoding actually
|
|
66
|
+
run here. The
|
|
67
|
+
wizard makes a small, audio-free copy for Gemini to watch, then measures the
|
|
68
|
+
segments Gemini proposes with `blockdetect`, `blurdetect`, `signalstats`,
|
|
69
|
+
`hqdn3d`, and `psnr`. `scdet` is used by legacy shot detection, not wizard
|
|
70
|
+
segmentation. `libvmaf` is optional for offline regression.
|
|
71
|
+
- A Gemini API key from [Google AI Studio](https://aistudio.google.com/) for the wizard
|
|
72
|
+
- For the upscale itself: an NVIDIA GPU with CUDA, PyTorch, spandrel, and the
|
|
73
|
+
SPAN weights file (see [GPU setup](#gpu-setup))
|
|
74
|
+
|
|
75
|
+
## Quick start
|
|
76
|
+
|
|
77
|
+
Releases are on PyPI as `ancient-processor`; the command it installs is
|
|
78
|
+
still `super-processor`:
|
|
79
|
+
|
|
80
|
+
```bash
|
|
81
|
+
python -m pip install ancient-processor # add [gpu] for the upscale
|
|
82
|
+
```
|
|
83
|
+
|
|
84
|
+
From a checkout:
|
|
85
|
+
|
|
86
|
+
```bash
|
|
87
|
+
python -m venv .venv
|
|
88
|
+
source .venv/bin/activate
|
|
89
|
+
python -m pip install -e ".[dev]"
|
|
90
|
+
|
|
91
|
+
export GEMINI_API_KEY=your_key # or paste it in the wizard setup screen
|
|
92
|
+
super-processor doctor
|
|
93
|
+
super-processor review
|
|
94
|
+
```
|
|
95
|
+
|
|
96
|
+
The wizard walks through Gemini setup, file pick (up to 1080p and 30 minutes),
|
|
97
|
+
segment review, a before/after preview of each segment in turn, a SPAN
|
|
98
|
+
upscale on an NVIDIA GPU, and a result screen where a note re-runs one time
|
|
99
|
+
range.
|
|
100
|
+
|
|
101
|
+
The wizard and restoration entry points currently require verified constant
|
|
102
|
+
frame rate (CFR). Before a job is created or restoration starts, a CPU decode
|
|
103
|
+
checks the complete presentation timeline, not just reported frame-rate
|
|
104
|
+
metadata. VFR, nonuniform timing, or timing that cannot be verified is refused
|
|
105
|
+
before GPU work or replacing an existing output. There is no automatic
|
|
106
|
+
normalization: external CFR conversion may change frames/timing. This is an
|
|
107
|
+
interim safety limit, not VFR support.
|
|
108
|
+
|
|
109
|
+
Without a GPU, the wizard works up to the Segments screen (the still grid and
|
|
110
|
+
Gemini's segments). The segment previews and the upscale need SPAN on CUDA.
|
|
111
|
+
|
|
112
|
+
## GPU setup
|
|
113
|
+
|
|
114
|
+
Install the GPU extra (PyTorch and spandrel, which loads the SPAN network):
|
|
115
|
+
|
|
116
|
+
```bash
|
|
117
|
+
python -m pip install -e ".[dev,gpu]"
|
|
118
|
+
```
|
|
119
|
+
|
|
120
|
+
Download the weights, `2xLiveActionV1_SPAN_490000.pth` (about 9 MB), to where
|
|
121
|
+
the wizard looks for them, `.dogfood/models/SPAN/` under the folder you run
|
|
122
|
+
from. The link is pinned to the commit that relicensed them as Apache-2.0
|
|
123
|
+
(see [THIRD_PARTY_NOTICES.md](THIRD_PARTY_NOTICES.md)):
|
|
124
|
+
|
|
125
|
+
```bash
|
|
126
|
+
mkdir -p .dogfood/models/SPAN
|
|
127
|
+
curl -L -o .dogfood/models/SPAN/2xLiveActionV1_SPAN_490000.pth \
|
|
128
|
+
https://github.com/jcj83429/upscaling/raw/5d8cdd2e17750b64be39ccab3a8763d91128fe15/2xLiveActionV1_SPAN/2xLiveActionV1_SPAN_490000.pth
|
|
129
|
+
sha256sum .dogfood/models/SPAN/2xLiveActionV1_SPAN_490000.pth
|
|
130
|
+
# 8b166c75831ea7f694d9058ee9c8df8148af8cc1d2b57e69e6581b15cab572f7
|
|
131
|
+
```
|
|
132
|
+
|
|
133
|
+
`SUPER_PROCESSOR_SPAN_WEIGHTS=/path/to/file.pth` reads the weights from
|
|
134
|
+
another place; `SUPER_PROCESSOR_MODELS` moves the whole models folder (the
|
|
135
|
+
file is then expected in its `SPAN/` folder). The wizard checks the file's
|
|
136
|
+
sha256 before it loads it and names the expected one if it differs. Then
|
|
137
|
+
check each piece:
|
|
138
|
+
|
|
139
|
+
```bash
|
|
140
|
+
python -c "import torch; print('CUDA:', torch.cuda.is_available())"
|
|
141
|
+
python -c "from super_processor.span_sr import resolve_span_weights as r; from super_processor.upscale import models_dir as m; print(r(m()))"
|
|
142
|
+
super-processor doctor
|
|
143
|
+
ffmpeg -hide_banner -filters | grep -E ' (blockdetect|blurdetect|signalstats|hqdn3d|psnr) '
|
|
144
|
+
```
|
|
145
|
+
|
|
146
|
+
The first should print `CUDA: True`, the second the weights file (it raises
|
|
147
|
+
with the expected path or hash otherwise). Doctor should report required tools
|
|
148
|
+
ready, and the final command should show all five wizard measurement filters.
|
|
149
|
+
Legacy shot detection additionally needs `scdet`.
|
|
150
|
+
|
|
151
|
+
SPAN sees one frame at a time and the frames go to the GPU in small batches,
|
|
152
|
+
so GPU memory follows the batch, not the clip: about 4-9 GB for 1440x1080 to
|
|
153
|
+
2880x2160. Each clip streams in one pass with nearly all of the work on the
|
|
154
|
+
GPU. FFmpeg decodes the source with NVDEC (`-hwaccel cuda`; codecs NVDEC
|
|
155
|
+
cannot read fall back to software decoding) and applies the look's clean
|
|
156
|
+
filter. On the GPU the frames go from YUV to RGB, through SPAN and the
|
|
157
|
+
strength blend with a lanczos upscale, and back to YUV, and `h264_nvenc`
|
|
158
|
+
encodes them from a pipe, so nothing is written to scratch disk. Only a
|
|
159
|
+
look's grade and grain, when it has them, still run on the CPU. On an RTX
|
|
160
|
+
5090, 1500 frames of 1440x1080 to 2880x2160 ran at 26.8 frames a second end
|
|
161
|
+
to end with 0.054 CPU-seconds per frame, and the CPU was at 76 C after the
|
|
162
|
+
minute. The previous stream blended and encoded with x264 on the CPU: 22.2
|
|
163
|
+
frames a second at 0.43 CPU-seconds per frame (about 9.5 of 32 threads, half
|
|
164
|
+
of it in x264), and a 20-minute film took the CPU to 95 C. Without NVENC
|
|
165
|
+
(Macs, AMD GPUs, CI) the delivery encode falls back to `libx264`. If the
|
|
166
|
+
GPU runs out of memory, the upscale retries with smaller batches, down to one
|
|
167
|
+
frame, and then stops with advice to close other GPU programs.
|
|
168
|
+
|
|
169
|
+
### Why SPAN and not FlashVSR
|
|
170
|
+
|
|
171
|
+
Up to 3.7.0 the upscaler was FlashVSR v1.1 (or FlashVSR+). A bake-off on an
|
|
172
|
+
RTX 5090 found that FlashVSR's output, blended at strength 0.5 as the wizard
|
|
173
|
+
uses it, was almost the same as plain SPAN 2x on SD footage (consistency 41.9
|
|
174
|
+
against 41.5 dB, detail 32.1 against 32.6, added shimmer +0.04 against +0.07),
|
|
175
|
+
but it took 97 minutes against 3 minutes for a 20-minute film and 26 GB
|
|
176
|
+
against 2 GB of GPU memory. On 1440x1080, FlashVSR did not fit a 32 GB GPU at
|
|
177
|
+
all, which is why it was limited to sources up to 1280 px. Recent jobs used
|
|
178
|
+
strength 0.5 or less for 85% of segments, never above 0.7, and never 4x. SPAN
|
|
179
|
+
replaced it outright, with 4x and the system-RAM fallback.
|
|
180
|
+
|
|
181
|
+
### Output size
|
|
182
|
+
|
|
183
|
+
While it proposes segments, Gemini also picks the output size for the whole
|
|
184
|
+
video and says why: 1x (keep the size, no upscaling) or 2x (the default). Both
|
|
185
|
+
are offered for every source the wizard accepts (up to 1080p). The Segments
|
|
186
|
+
screen shows Gemini's choice and reason, and the Ready screen shows the reason
|
|
187
|
+
while that choice stands; you can still change the size there. Segment
|
|
188
|
+
previews render at the plan's scale, and a 1x job never runs SPAN for its
|
|
189
|
+
previews.
|
|
190
|
+
|
|
191
|
+
### Progress and time left
|
|
192
|
+
|
|
193
|
+
Both interfaces show a bar for the whole job ("N% of the job") and a line for
|
|
194
|
+
the running step: its percentage, what it is doing, and "about X left", or
|
|
195
|
+
"X so far" while Gemini is thinking. Segmenting is the first 8% of the job,
|
|
196
|
+
the previews the next 30% (split evenly per segment), and the render the
|
|
197
|
+
rest; a render from a result note counts 0–100% on its own. The time left
|
|
198
|
+
ignores time before the step's first report, such as loading SPAN, and
|
|
199
|
+
follows both the step's average speed and its recent speed. In the render,
|
|
200
|
+
an upscaled frame counts more than a frame that is only resized.
|
|
201
|
+
|
|
202
|
+
## Testing on your own footage
|
|
203
|
+
|
|
204
|
+
```bash
|
|
205
|
+
export GEMINI_API_KEY=your_key
|
|
206
|
+
mkdir -p .dogfood/jobs
|
|
207
|
+
super-processor --jobs-dir .dogfood/jobs review
|
|
208
|
+
```
|
|
209
|
+
|
|
210
|
+
Open the printed URL and go pick → segments → previews → upscale → result. Short
|
|
211
|
+
clips (30–60 seconds) with one clear problem each are the most useful: blocky
|
|
212
|
+
web video, noisy low light, old SD footage, a face close-up, and a shot with
|
|
213
|
+
a sign or text. On each screen, note where the tool was wrong:
|
|
214
|
+
|
|
215
|
+
- **Segments**: cuts in the wrong place, or contents (faces, hands, text)
|
|
216
|
+
missed (`segments.json`; `shots.json` holds the measurements).
|
|
217
|
+
- **Previews**: waxy skin, changed faces, garbled text, or crawling texture in a
|
|
218
|
+
preview; face and text boxes in the wrong place; a shimmer number that does
|
|
219
|
+
not match what you see (`shot_plan.json`).
|
|
220
|
+
- **Result**: seam warnings with no visible seam, or visible seams with no
|
|
221
|
+
warning (`render_report.json`).
|
|
222
|
+
|
|
223
|
+
Each job folder under `.dogfood/jobs/` keeps those files, the previews, and
|
|
224
|
+
`gemini_chat.json`. The thresholds they report against are first guesses from
|
|
225
|
+
synthetic clips and are meant to be tuned from real footage:
|
|
226
|
+
|
|
227
|
+
| What | Where | Value |
|
|
228
|
+
|---|---|---|
|
|
229
|
+
| Problem hints | `shots.py` | blocky ≥ 18, soft ≥ 6, noisy ≥ 2, dark < 0.25, bright > 0.75, flat < 0.35 |
|
|
230
|
+
| Shimmer | `stability.py` `SHIMMER_LIMIT` | 1.6 |
|
|
231
|
+
| Seams | `stability.py` `SEAM_LIMIT` | 2.5 |
|
|
232
|
+
| Protected regions | `regions.py` `REGION_STRENGTH` | faces and hands 0.3, text 0.15 |
|
|
233
|
+
|
|
234
|
+
Legacy single-grade and segment CLI commands remain available:
|
|
235
|
+
|
|
236
|
+
```bash
|
|
237
|
+
super-processor diagnose ./clip.mp4 --max-size-mb 80
|
|
238
|
+
super-processor plan JOB_ID --instruction "less denoise"
|
|
239
|
+
super-processor preview JOB_ID --open auto
|
|
240
|
+
super-processor apply JOB_ID --approve
|
|
241
|
+
super-processor show JOB_ID
|
|
242
|
+
```
|
|
243
|
+
|
|
244
|
+
See [docs/ui-wizard.md](docs/ui-wizard.md),
|
|
245
|
+
[docs/system-design.md](docs/system-design.md), the
|
|
246
|
+
[ordered release plan](docs/release-plan.md), and the
|
|
247
|
+
[acceptance matrix and pending evidence](docs/acceptance.md).
|
|
248
|
+
|
|
249
|
+
The `web/` app is a React shell for that wizard. GitHub Pages can host it as
|
|
250
|
+
a static preview. The GPU upscale still runs only from `super-processor review`.
|
|
251
|
+
Start `super-processor review` first and copy its printed backend URL: the
|
|
252
|
+
server chooses an ephemeral port, not necessarily `45143`. In `web/`, install
|
|
253
|
+
with `npm ci`, then run `VITE_PROXY_TARGET=http://127.0.0.1:PRINTED_PORT npm run dev`
|
|
254
|
+
(replace `PRINTED_PORT` with the actual port). This proxies API, uploads, stills,
|
|
255
|
+
previews, and output. For a built app, use the same variable with
|
|
256
|
+
`npm run preview` after `npm run build`. The fallback proxy target
|
|
257
|
+
`http://127.0.0.1:45143` works only if that is the backend's actual address.
|
|
258
|
+
|
|
259
|
+
Alternatively, enter the printed backend URL in the local shell's connection
|
|
260
|
+
form or open `http://localhost:5173/?api=http://127.0.0.1:PRINTED_PORT` for a
|
|
261
|
+
direct connection (preview uses port `4173`). See
|
|
262
|
+
[frontend connection and recovery](docs/ui-wizard.md#react-shell-connection-and-recovery)
|
|
263
|
+
for CORS, HTTPS restrictions, and retry behavior. Offline frontend regressions:
|
|
264
|
+
`cd web && npm test && npm run build` (no API key, GPU, or browser required).
|
|
265
|
+
|
|
266
|
+
## Development checks
|
|
267
|
+
|
|
268
|
+
```bash
|
|
269
|
+
python -m ruff format --check .
|
|
270
|
+
python -m ruff check .
|
|
271
|
+
python -m mypy
|
|
272
|
+
python -m pytest -m 'not gemini_live'
|
|
273
|
+
python -m build
|
|
274
|
+
super-processor self-test
|
|
275
|
+
```
|
|
276
|
+
|
|
277
|
+
Routine tests exclude live Gemini calls even when `GEMINI_API_KEY` is set.
|
|
278
|
+
Live tests are a separate, explicit opt-in: with a test key configured, run
|
|
279
|
+
`python -m pytest -m gemini_live --no-cov`. These contact Google, send test
|
|
280
|
+
prompts/context, and consume API quota (potentially incurring charges).
|
|
281
|
+
Wizard use also uploads an audio-free copy of the whole video (long edge at most
|
|
282
|
+
960 px) to Google's File API, where it is kept about 48 hours, and sends
|
|
283
|
+
selected stills, notes, and chat context; do not use sensitive footage without
|
|
284
|
+
permission. See [CONTRIBUTING.md](CONTRIBUTING.md) for testing
|
|
285
|
+
and packaging checks, and [CHANGELOG.md](CHANGELOG.md) for changes.
|
|
@@ -0,0 +1,47 @@
|
|
|
1
|
+
ancient_processor-4.0.1.dist-info/licenses/LICENSE,sha256=-bCLq3SN_Dh45pYRVh26HbLnE_Ar_Q-UyMQ1uFmPuz4,1085
|
|
2
|
+
super_processor/__init__.py,sha256=SsuutK4rRjqosBdvSDzYaBPrQJ6vf8JoQ-48Uh5Hj7w,1116
|
|
3
|
+
super_processor/__main__.py,sha256=RUTBxsbjfbcYDQI70JGpPDTmKD72qQfDcM1YHeWzPOo,102
|
|
4
|
+
super_processor/_core.cp314-win_amd64.pyd,sha256=RPBIJBubg-2l3WXXpcYRqkXbD2ziyG8PMG-vNM-mrwA,145920
|
|
5
|
+
super_processor/_core.pyi,sha256=dIN_V6JThHgEonvC_oxIZFenjJXRrYmFr9mK-D6UcQs,900
|
|
6
|
+
super_processor/_version.py,sha256=YfFyKSm-EHvFlwAibk5zvWm1-rjNZ2TF39q0eAbQwa4,552
|
|
7
|
+
super_processor/cli.py,sha256=QhYEcX2XLp5NcuYa1npN2b8S4mJOt2uUHCY-GDewnMM,38923
|
|
8
|
+
super_processor/diagnose.py,sha256=omuwfyfmXkf1NFnnNwl0zC7pWcvYxhl4N4oHE5ATZbo,6321
|
|
9
|
+
super_processor/doctor.py,sha256=KVdIa1Z1kDtgJNg3BPdDxOPLgsAGZf_xCya257vi630,9231
|
|
10
|
+
super_processor/encoders.py,sha256=V6ePZkCQ4_6lWl2PZSoF0oLjZ--VCfSRI21Qq3ILaEc,2955
|
|
11
|
+
super_processor/estimators.py,sha256=WstQhLzS3v8iRKRZj3tBg3MseNG8ViMxGqBxGcfPV0o,19688
|
|
12
|
+
super_processor/file_dialog.py,sha256=SjqiJisfCS5HYUIV5epxD601SbRwpfS0jWzoPVWaL74,3932
|
|
13
|
+
super_processor/gemini.py,sha256=qxmoNeVZ8_PYJLF4gGn26cUUtg6qRvrP0qxAzh0Mmx4,65773
|
|
14
|
+
super_processor/gpu_heat.py,sha256=AlioGoRqHWGSQlJ9jdTmybONp6WOuIU7WsYcQXZ42Ho,5646
|
|
15
|
+
super_processor/jobs.py,sha256=yMLXykjFn5xGzNzBR6jIhb7GO1vt2YX6wrXwOJkhd3Q,9281
|
|
16
|
+
super_processor/look.py,sha256=EGY5iaPn6D0kRUYfY92D_qQn7vVHlLKJwSFOiz4dIZ8,4987
|
|
17
|
+
super_processor/models.py,sha256=E6LmOSUOrunb98Btl5XYhL9grICsW9OAHCjtck9bNsE,9196
|
|
18
|
+
super_processor/plan.py,sha256=lxENK4PNUvfPI_49M2EOGp429szZc78mB2Jold9KOEw,5056
|
|
19
|
+
super_processor/plans.py,sha256=lgmblK2iN3A295CnFK8ZH3wQ8YCfq4McHeEalsE_OwY,12978
|
|
20
|
+
super_processor/preview.py,sha256=9RKud63IsLIwR5kdmsQX2-59sIXvpPcHe1kk7ryjExc,7442
|
|
21
|
+
super_processor/probe.py,sha256=TsGOaGoLxn3WR1Dt9PMYNkHF5LVJCRo_RlozEoWUa2Y,16583
|
|
22
|
+
super_processor/progress.py,sha256=gqy5vJDolgQx-K5RmmgiyrEvfUhGvs89HYk3L4rN7I4,4156
|
|
23
|
+
super_processor/py.typed,sha256=WqcInkQHOl79Ubpr3txasR-vjkpDYxhXXlxkUzumLNo,44
|
|
24
|
+
super_processor/qa.py,sha256=M-UxQnC48aF0RBSZCOQUV_96Lhsjp7qhY0wyWC5Eqzg,4751
|
|
25
|
+
super_processor/recipe.py,sha256=9eG3wPSG-KRixUFu40oCr6l2un2vfqf-U1zC5QnTECs,9882
|
|
26
|
+
super_processor/reframe.py,sha256=k1cNHnKexJXtdzBrgyc16-igVk2pWNFoVT7BUgZCRTo,10911
|
|
27
|
+
super_processor/regions.py,sha256=lVpZ8FIibykmKq73smMsjhIBUqZYBqMQ94ReYYcAR0M,6045
|
|
28
|
+
super_processor/render.py,sha256=BgtbKSFrmWW2OAfz2XP2QGOHRNf19h8BXvrUjwCzSls,4561
|
|
29
|
+
super_processor/review.py,sha256=ne3buqa-HaFH_8u2JxXQiC1IqgIFCBODKkaEqtbYzMQ,34206
|
|
30
|
+
super_processor/segment_loop.py,sha256=Rbe5QVSvfjFNS2zWCAZhW9F1aeajW9U0R0bx7M5GZLQ,17677
|
|
31
|
+
super_processor/segments.py,sha256=IZL-4f2ZBC8vnPlmBzFm_0zM4XSCQ0lzFqhopviuEKg,30766
|
|
32
|
+
super_processor/shot_plan.py,sha256=D_8Fj1GUip2sZXOSQiS5jhFKkt2pH3GriWP8fgSxQIY,29246
|
|
33
|
+
super_processor/shots.py,sha256=oCs413tN0zHCyaYd0kpFUzOAn-HGgZYXRyvoF-naXTQ,18921
|
|
34
|
+
super_processor/span_sr.py,sha256=E9jkYcwhUKqMUEn24hvyWvGNEEK7QKKyACsN8A75fyQ,16580
|
|
35
|
+
super_processor/stability.py,sha256=AZsYLEBu6_S0NPVZRshekRk9k_8ALbXPI9zpPmoIaRQ,9836
|
|
36
|
+
super_processor/templates.py,sha256=xZp8yzD6NZep8WN4a4ofQM6wrNyvaYdo7d9BMNi80dw,12337
|
|
37
|
+
super_processor/treatments.py,sha256=Knww2EnVdw-lyxl0Io3RqNcMTaw7ybV6S3OAIu-eoO0,10756
|
|
38
|
+
super_processor/upscale.py,sha256=2YJ-zNFAXVOlgqjWq5Bn6axdZDn7AS4KFJSX7fpNOII,86771
|
|
39
|
+
super_processor/validator.py,sha256=jV9G3DCh22W5YJuBsBAZCwKW6unk3KUG9f5FRa2H9o4,11333
|
|
40
|
+
super_processor/wizard.py,sha256=MnbSCMIscLDERVTKgWs2clWLV7TJmkSk7_tFyfM03HM,76239
|
|
41
|
+
super_processor/wizard_pages.py,sha256=rMjCBZUg9xovoS0Crhg06VKMllNjlvjau6DJdiY4RWQ,39319
|
|
42
|
+
super_processor/worker.py,sha256=s9_kEiiDJmjL4YFIJkFS1uc8Y3WNG5aGfJW3M9uiDCg,10136
|
|
43
|
+
ancient_processor-4.0.1.dist-info/METADATA,sha256=ybmjEikkp01Ay2uQk1zDkcoCnaIt1NdoDj3Nnq3t0Jk,13978
|
|
44
|
+
ancient_processor-4.0.1.dist-info/WHEEL,sha256=GjNoWUKTwVqYOh-GAeOqmWs8K5Mlz_Ym2F39m91XU8Q,101
|
|
45
|
+
ancient_processor-4.0.1.dist-info/entry_points.txt,sha256=VT0Wx91Z2Q3NeLQ3ul6GjXktzXWSIgSUyGawrke8oI8,61
|
|
46
|
+
ancient_processor-4.0.1.dist-info/top_level.txt,sha256=R2LIIKnt-3N0KCb4FCty60inrXuTgNuDwMgkTuDse7U,16
|
|
47
|
+
ancient_processor-4.0.1.dist-info/RECORD,,
|
|
@@ -0,0 +1,21 @@
|
|
|
1
|
+
MIT License
|
|
2
|
+
|
|
3
|
+
Copyright (c) 2026 Super Processor contributors
|
|
4
|
+
|
|
5
|
+
Permission is hereby granted, free of charge, to any person obtaining a copy
|
|
6
|
+
of this software and associated documentation files (the "Software"), to deal
|
|
7
|
+
in the Software without restriction, including without limitation the rights
|
|
8
|
+
to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
|
|
9
|
+
copies of the Software, and to permit persons to whom the Software is
|
|
10
|
+
furnished to do so, subject to the following conditions:
|
|
11
|
+
|
|
12
|
+
The above copyright notice and this permission notice shall be included in all
|
|
13
|
+
copies or substantial portions of the Software.
|
|
14
|
+
|
|
15
|
+
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
|
|
16
|
+
IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
|
|
17
|
+
FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
|
|
18
|
+
AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
|
|
19
|
+
LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
|
|
20
|
+
OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
|
|
21
|
+
SOFTWARE.
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
super_processor
|
|
@@ -0,0 +1,53 @@
|
|
|
1
|
+
"""Public package interface for Super Processor."""
|
|
2
|
+
|
|
3
|
+
from importlib.metadata import PackageNotFoundError, version
|
|
4
|
+
|
|
5
|
+
from ._core import (
|
|
6
|
+
histogram_u8,
|
|
7
|
+
mean_luma,
|
|
8
|
+
percentile_u8,
|
|
9
|
+
sad_u8,
|
|
10
|
+
sum_squares,
|
|
11
|
+
variance_u8,
|
|
12
|
+
)
|
|
13
|
+
from .jobs import (
|
|
14
|
+
InvalidTransitionError,
|
|
15
|
+
JobError,
|
|
16
|
+
JobManifest,
|
|
17
|
+
JobNotFoundError,
|
|
18
|
+
JobState,
|
|
19
|
+
JobStore,
|
|
20
|
+
default_jobs_root,
|
|
21
|
+
)
|
|
22
|
+
from .probe import MediaFacts, ProbeError, StreamFacts, probe_file
|
|
23
|
+
from .recipe import OpName, Recipe, RecipeError, empty_recipe
|
|
24
|
+
|
|
25
|
+
try:
|
|
26
|
+
__version__ = version("ancient-processor")
|
|
27
|
+
except PackageNotFoundError: # pragma: no cover - source tree without installation
|
|
28
|
+
__version__ = "0+unknown"
|
|
29
|
+
|
|
30
|
+
__all__ = [
|
|
31
|
+
"InvalidTransitionError",
|
|
32
|
+
"JobError",
|
|
33
|
+
"JobManifest",
|
|
34
|
+
"JobNotFoundError",
|
|
35
|
+
"JobState",
|
|
36
|
+
"JobStore",
|
|
37
|
+
"MediaFacts",
|
|
38
|
+
"OpName",
|
|
39
|
+
"ProbeError",
|
|
40
|
+
"Recipe",
|
|
41
|
+
"RecipeError",
|
|
42
|
+
"StreamFacts",
|
|
43
|
+
"__version__",
|
|
44
|
+
"default_jobs_root",
|
|
45
|
+
"empty_recipe",
|
|
46
|
+
"histogram_u8",
|
|
47
|
+
"mean_luma",
|
|
48
|
+
"percentile_u8",
|
|
49
|
+
"probe_file",
|
|
50
|
+
"sad_u8",
|
|
51
|
+
"sum_squares",
|
|
52
|
+
"variance_u8",
|
|
53
|
+
]
|
|
Binary file
|
|
@@ -0,0 +1,23 @@
|
|
|
1
|
+
def sum_squares(count: int) -> int:
|
|
2
|
+
"""Return the sum of squared integers below ``count``."""
|
|
3
|
+
|
|
4
|
+
def mean_luma(samples: bytes | bytearray | memoryview) -> float:
|
|
5
|
+
"""Return the arithmetic mean of a contiguous 8-bit luma buffer."""
|
|
6
|
+
|
|
7
|
+
def histogram_u8(
|
|
8
|
+
samples: bytes | bytearray | memoryview,
|
|
9
|
+
out_256: bytes | bytearray | memoryview,
|
|
10
|
+
) -> None:
|
|
11
|
+
"""Fill a 256-bin histogram for an 8-bit buffer."""
|
|
12
|
+
|
|
13
|
+
def percentile_u8(samples: bytes | bytearray | memoryview, percentile: float) -> float:
|
|
14
|
+
"""Return an approximate percentile from an 8-bit buffer via histogram."""
|
|
15
|
+
|
|
16
|
+
def variance_u8(samples: bytes | bytearray | memoryview) -> float:
|
|
17
|
+
"""Return the population variance of an 8-bit buffer."""
|
|
18
|
+
|
|
19
|
+
def sad_u8(
|
|
20
|
+
left: bytes | bytearray | memoryview,
|
|
21
|
+
right: bytes | bytearray | memoryview,
|
|
22
|
+
) -> int:
|
|
23
|
+
"""Return the sum of absolute differences between two equal-length buffers."""
|
|
@@ -0,0 +1,24 @@
|
|
|
1
|
+
# file generated by vcs-versioning
|
|
2
|
+
# don't change, don't track in version control
|
|
3
|
+
from __future__ import annotations
|
|
4
|
+
|
|
5
|
+
__all__ = [
|
|
6
|
+
"__version__",
|
|
7
|
+
"__version_tuple__",
|
|
8
|
+
"version",
|
|
9
|
+
"version_tuple",
|
|
10
|
+
"__commit_id__",
|
|
11
|
+
"commit_id",
|
|
12
|
+
]
|
|
13
|
+
|
|
14
|
+
version: str
|
|
15
|
+
__version__: str
|
|
16
|
+
__version_tuple__: tuple[int | str, ...]
|
|
17
|
+
version_tuple: tuple[int | str, ...]
|
|
18
|
+
commit_id: str | None
|
|
19
|
+
__commit_id__: str | None
|
|
20
|
+
|
|
21
|
+
__version__ = version = '4.0.1'
|
|
22
|
+
__version_tuple__ = version_tuple = (4, 0, 1)
|
|
23
|
+
|
|
24
|
+
__commit_id__ = commit_id = 'g516286b08'
|