@@ -230,30 +230,106 @@ const Documentation = () => {
230230 requirements below before cloning the repository.
231231 </ p >
232232 </ div >
233- < div className = "grid gap-6 md:grid-cols-[1fr_minmax(0,1.2fr)]" >
233+
234+ { /* Three-column responsive grid */ }
235+ < div className = "grid gap-6 md:grid-cols-3" >
236+ { /* MacOS Box */ }
234237 < div className = "rounded-xl bg-card border p-6 space-y-4" >
235- < h3 className = "text-xl font-semibold" > Requirements</ h3 >
238+ < h3 className = "text-xl font-semibold" > Requirements — macOS </ h3 >
236239 < ul className = "list-disc pl-5 space-y-2 text-muted-foreground" >
237- < li > Python ≥ 3.11</ li >
238- < li > FFmpeg</ li >
239- < li > OpenCV</ li >
240- < li > TensorFlow</ li >
241- < li > PyQt5</ li >
242- < li > loguru</ li >
243- < li > tqdm</ li >
244- < li > psutil</ li >
240+ < li > certifi==2025.10.5</ li >
241+ < li > charset-normalizer==3.4.4</ li >
242+ < li > colorlog==6.10.1</ li >
243+ < li > contourpy==1.3.2</ li >
244+ < li > cycler==0.12.1</ li >
245+ < li > fonttools==4.60.1</ li >
246+ < li > idna==3.11</ li >
247+ < li > kiwisolver==1.4.9</ li >
248+ < li > loguru==0.7.3</ li >
249+ < li > matplotlib==3.10.7</ li >
250+ < li > numpy==2.2.6</ li >
251+ < li > opencv-python==4.12.0.88</ li >
252+ < li > packaging==25.0</ li >
253+ < li > Pillow==12.0.0</ li >
254+ < li > psutil==7.1.1</ li >
255+ < li > pyparsing==3.2.5</ li >
256+ < li > PyQt5==5.15.11</ li >
257+ < li > PyQt5-Qt5==5.15.17</ li >
258+ < li > PyQt5_sip==12.17.1</ li >
259+ < li > python-dateutil==2.9.0.post0</ li >
260+ < li > requests==2.32.5</ li >
261+ < li > six==1.17.0</ li >
262+ < li > tqdm==4.67.1</ li >
263+ < li > urllib3==2.5.0</ li >
264+ < li > vidgear==0.3.3</ li >
245265 </ ul >
246266 </ div >
267+
268+ { /* Windows Box */ }
269+ < div className = "rounded-xl bg-card border p-6 space-y-4" >
270+ < h3 className = "text-xl font-semibold" > Requirements — Windows</ h3 >
271+ < p className = "text-muted-foreground text-sm" > Requires Python 3.10 on Windows</ p >
272+ < ul className = "list-disc pl-5 space-y-2 text-muted-foreground" >
273+ < li > tensorflow-cpu==2.10.1</ li >
274+ < li > numpy>=1.23,<2.0</ li >
275+ < li > opencv-python<4.9</ li >
276+ < li > PyQt5==5.15.11</ li >
277+ < li > vidgear==0.3.3</ li >
278+ < li > tqdm>=4.66</ li >
279+ < li > requests>=2.31</ li >
280+ < li > protobuf==3.19.6</ li >
281+ < li > tensorboard<2.11</ li >
282+ < li > pillow>=9.4</ li >
283+ < li > loguru>=0.7</ li >
284+ < li > colorlog>=6.8</ li >
285+ < li > matplotlib==3.10.7</ li >
286+ < li > psutil==7.1.1</ li >
287+ </ ul >
288+ </ div >
289+
290+ { /* From Source Box */ }
247291 < div className = "rounded-xl bg-card border p-6 space-y-4" >
248292 < h3 className = "text-xl font-semibold" > From Source</ h3 >
249293 < p className = "text-muted-foreground" >
250- Use the following commands to clone, install dependencies, and launch the application.
294+ Follow these steps to clone the repository, install dependencies, download model weights, and
295+ set up FFmpeg before launching Endoshare.
251296 </ p >
252297 < pre className = "bg-muted rounded-lg p-4 text-sm overflow-x-auto" >
253- < code > { `git clone https://github.com/your-org/Endoshare.git
254- cd Endoshare
255- pip install -r requirements.txt
256- python app.py` } </ code >
298+ < code > { `# 1. Clone the repository
299+ git clone https://github.com/CAMMA-public/Endoshare_code.git
300+ cd Endoshare
301+
302+ # 2. Install Python dependencies
303+ pip install -r requirements.txt
304+
305+ # 3. Download model weights (OOBNet)
306+ # Create the checkpoint directory and download weights into it
307+ cd endoshare/resources
308+ mkdir -p ckpt
309+ cd ckpt
310+ wget https://s3.unistra.fr/camma_public/github/oobnet_detection/ckpt/oobnet_weights.h5
311+ cd ../../../
312+
313+ # 4. Install FFmpeg binaries
314+ # The FFmpeg executable must be placed inside: endoshare/Externals/ffmpeg/
315+
316+ # ── macOS / Linux ────────────────────────────────
317+ # Create the folder and download the static build
318+ mkdir -p endoshare/Externals/ffmpeg
319+ cd endoshare/Externals/ffmpeg
320+ wget https://johnvansickle.com/ffmpeg/releases/ffmpeg-release-amd64-static.tar.xz
321+ tar -xf ffmpeg-release-amd64-static.tar.xz
322+ mv ffmpeg-*-static/ffmpeg .
323+ cd ../../../..
324+
325+ # ── Windows ─────────────────────────────────────
326+ # 1. Go to https://www.gyan.dev/ffmpeg/builds/
327+ # 2. Download "ffmpeg-git-full.7z"
328+ # 3. Extract it and copy the ffmpeg.exe binary into:
329+ # endoshare\\Externals\\ffmpeg\\ffmpeg.exe
330+
331+ # 5. Launch Endoshare
332+ python main.py` } </ code >
257333 </ pre >
258334 </ div >
259335 </ div >
@@ -401,7 +477,7 @@ python app.py`}</code>
401477
402478 < section className = "rounded-2xl border bg-muted/30 p-8 space-y-4 text-center" >
403479 < h2 className = "text-2xl font-semibold" > Important Disclaimer</ h2 >
404- < p className = "text-muted-foreground max-w-3xl mx-auto" >
480+ < p className = "text-muted-foreground max-w-4xl mx-auto" >
405481 Endoshare is research and educational software. It is not a medical device and is not certified for clinical or
406482 diagnostic use. Manual review of outputs is recommended to ensure privacy assurance before sharing any content.
407483 </ p >
0 commit comments