Rebuild voSINT v2: token-free Playwright adapters, modular CLI, Pinterest-first default

.env.example

@@ -0,0 +1,5 @@
+# voSINT v2 is token-free by default.
+# Optional flags only.
+VOSINT_MODE=deep
+VOSINT_HEADFUL=false
+VOSINT_CASES_DIR=cases

README.md

@@ -1,55 +1,74 @@
-<p align="center">
- <img src="Results/logo.png" width="500">
-</p>
-# voSINT: Video Reverse Search OSINT Tool
+# voSINT v2
 
-## Description
-voSINT is an Open Source Intelligence (OSINT) tool designed for video reverse search. It enables users to trace the digital footprint of a video across the internet. By listing the results in descending order, voSINT reveals where a video first appeared and its subsequent occurrences online. This tool is invaluable for cybersecurity experts, digital forensics analysts, and anyone interested in the origin and spread of digital content.
+Token-free reverse-video OSINT workflow focused on **origin hunting** and **repost spread** analysis.
 
-Key Features:
-- Track video appearances online in descending order.
-- Generate approximate results, prioritizing data scope.
-- Beta version focused on user feedback and continuous improvement.
+## Highlights
+- No SerpApi key, no `config.ini` setup, no upload-to-host requirement.
+- Playwright provider adapters (default order): **Pinterest**, Google Lens, Bing Visual, Yandex, TinEye.
+- Multi-frame extraction with quality scoring and timeline-first ranking.
+- OCR pivots, transcript pivots (optional local dependencies), and Video DNA artifact.
+- Case-folder outputs: HTML/JSON/CSV + raw/normalized artifacts.
+- Commands: `scan`, `diff`, `report`.
 
-## Installation Guide
-Navigate to the directory where you want to create your project.
-
-### Setting up a Virtual Environment
-Run the following command to create a virtual environment (replace 'venv' with your desired environment name):
+## Install
 ```bash
-python3 -m venv venv
+python -m venv .venv
+source .venv/bin/activate
+pip install -e .
+python -m playwright install chromium
 ```
-Activate the virtual environment:
+
+## CLI
+### Single scan
 ```bash
-source venv/bin/activate
+vosint scan video.mp4
 ```
-Install the required packages:
+
+### Batch scan
 ```bash
-pip install -r requirements.txt
+vosint scan ./videos --batch --mode deep
 ```
 
-## Usage Instructions
-For using the tool with a single video:
+### Compare videos
 ```bash
-python voSINT.py <video_path>
+vosint diff a.mp4 b.mp4
 ```
-For multiple videos in a directory:
+
+### Re-open a case report
 ```bash
-python voSINT.py <videos_dir>
+vosint report cases/case_YYYYMMDD_HHMMSS
 ```
 
-By creating and activating a virtual environment, you ensure that the installed packages and dependencies are isolated from your system's global Python environment, providing a clean and separate environment for your project.
-
-## API Key Configuration
-Before using voSINT, you need to obtain an API key from SerpApi.com. This key is essential for the tool to perform video reverse searches in Google and Yandex without dealing with CAPTCHA. Follow these steps to configure your API key:
-
-Visit SerpApi.com and sign up to receive an API key.
+## Modes
+- `fast`: few top frames, Pinterest + Google Lens + Bing.
+- `deep` (default): more frames, OCR+transcript, all default providers.
+- `stealth`: local-only extraction/pivots, no provider submission; emits manual query pack.
 
-Once you have your API key, open the config.ini file in the voSINT directory.
-
-Insert your API key in the designated section of config.ini.
-
-Ensure your API key is correctly saved in the configuration file to enable the full functionality of voSINT.
+## Common flags
+```bash
+--mode fast|deep|stealth
+--providers pinterest,google_lens,bing_visual,yandex,tineye
+--max-frames 8
+--ocr
+--transcribe
+--json --csv --html
+--keep-frames
+--no-browser
+--headful
+```
 
+## Output layout
+Each run writes:
+```
+cases/<case_id>/
+  input/
+  frames/
+  raw/
+  normalized/
+  report.html
+  report.json
+  timeline.csv
+```
 
-![](https://raw.githubusercontent.com/Meshall/voSINT/master/walkthrough.gif)
+## Privacy note
+Reverse-image providers receive submitted frames unless `--mode stealth` or `--no-browser` is used.

pyproject.toml

@@ -0,0 +1,26 @@
+[build-system]
+requires = ["setuptools>=68", "wheel"]
+build-backend = "setuptools.build_meta"
+
+[project]
+name = "vosint"
+version = "2.0.0"
+description = "Token-free video reverse-search and provenance OSINT toolkit"
+readme = "README.md"
+requires-python = ">=3.10"
+dependencies = [
+  "numpy>=1.25",
+  "playwright>=1.45",
+  "imageio>=2.34",
+  "pillow>=10.0",
+]
+
+[project.optional-dependencies]
+ocr = ["pytesseract>=0.3.10"]
+transcribe = ["openai-whisper>=20231117"]
+
+[project.scripts]
+vosint = "vosint.cli:main"
+
+[tool.setuptools.packages.find]
+include = ["vosint*"]

requirements.txt

@@ -1,9 +1,4 @@
-certifi==2023.7.22
-charset-normalizer==3.2.0
-google-search-results==2.4.2
-idna==3.4
-numpy==1.25.2
-opencv-python==4.8.0.76
-requests==2.31.0
-tqdm==4.66.1
-urllib3==2.0.4
+numpy>=1.25
+playwright>=1.45
+imageio>=2.34
+pillow>=10.0

tests/test_normalize.py

@@ -0,0 +1,13 @@
+from vosint.core.normalize import normalize_hits
+from vosint.models import Hit
+
+
+def test_normalize_merges_duplicates():
+    hits = [
+        Hit(engine="pinterest", frame_id="f1", url="HTTPS://Example.com/path/", title="a"),
+        Hit(engine="bing_visual", frame_id="f2", url="https://example.com/path", title="b"),
+    ]
+    merged = normalize_hits(hits)
+    assert len(merged) == 1
+    assert merged[0].domain == "example.com"
+    assert merged[0].support_engines == {"pinterest", "bing_visual"}

tests/test_timeline.py

@@ -0,0 +1,10 @@
+from vosint.core.timeline import rank_timeline
+from vosint.models import Hit
+
+
+def test_rank_timeline_known_dates_first():
+    a = Hit(engine="p", frame_id="1", url="https://a", date_raw="2021-01-01")
+    b = Hit(engine="p", frame_id="2", url="https://b")
+    a.date_parsed = __import__("datetime").datetime(2021, 1, 1)
+    out = rank_timeline([b, a])
+    assert out[0].url == "https://a"