feat(ml): add TA-Lib annotation generation and import workflow

Add complete workflow for using TA-Lib to bootstrap training data: - generate_talib_annotations.py: Python script to run TA-Lib CDL* functions and output span annotations in UI-compatible format - import_talib_annotations.ts: TypeScript script to import generated annotations into the UI database with auto-label-type creation - npm script 'import-annotations' for easy execution - TALIB_WORKFLOW.md: Comprehensive guide covering the full cycle: * Generate patterns with TA-Lib * Import into UI * Review and edit in browser * Export and train model * Compare predictions with TA-Lib detections * Iterate for improvement This enables the intended workflow: use TA-Lib for initial annotations, manually refine them, then train a model that learns from corrections.
2026-02-15 19:18:28 +01:00 · 2026-02-15 19:18:28 +01:00 · 847ff67986
commit 847ff67986
parent 228f70daf3
18 changed files with 5416 additions and 7 deletions
--- a/.env
+++ b/.env
@ -0,0 +1,9 @@
+NODE_ENV=production
+PORT=3000
+DATABASE_PATH=/app/data/candles.db
+
+# ML Inference Service Configuration
+INFERENCE_API_URL=http://localhost:8001
+INFERENCE_API_TIMEOUT=30000
+INFERENCE_BATCH_TIMEOUT=120000
+NEXT_PUBLIC_PREDICTIONS_ENABLED=true
--- a/DEPLOYMENT.md
+++ b/DEPLOYMENT.md
@ -196,7 +196,8 @@ sudo make install

 ```bash
 cd services/ml
-pip install -r requirements.txt
+uv sync
+#pip install -r requirements.txt
 ```

 #### 3. Setup PostgreSQL
@ -217,7 +218,7 @@ DVC is used for dataset versioning:

 ```bash
 cd services/ml
-dvc init
+dvc init #--subdir
 dvc remote add -d local /path/to/dvc-storage
 ```

--- a/TALIB_WORKFLOW.md
+++ b/TALIB_WORKFLOW.md
@ -0,0 +1,372 @@
+# TA-Lib Annotation Workflow
+
+This guide shows how to use TA-Lib to generate initial pattern annotations, edit them in the UI, and train a model.
+
+## Overview
+
+1. **Generate** - Run TA-Lib CDL* functions to detect patterns automatically
+2. **Import** - Import detected patterns into the UI database
+3. **Review & Edit** - View, correct, and refine annotations in the web UI
+4. **Train** - Export annotations and train your model
+5. **Iterate** - Get predictions, compare with TA-Lib, retrain
+
+## Step 1: Generate TA-Lib Annotations
+
+### 1.1 Prepare Your Data
+
+Export candles from your database:
+
+```bash
+# From host
+docker-compose exec candle-annotator sh -c "sqlite3 /app/data/candles.db -csv -header 'SELECT time, open, high, low, close, volume FROM candles ORDER BY time;'" > OHLCV.csv
+
+# Or copy your existing CSV
+cp your_data.csv OHLCV.csv
+```
+
+### 1.2 Run Pattern Detection
+
+Enter the ML service container and run the generator:
+
+```bash
+# Enter container
+docker-compose exec ml-service bash
+
+# Generate annotations (all patterns, perfect matches only)
+python generate_talib_annotations.py \
+  --input data/raw/OHLCV.csv \
+  --output talib_annotations.json
+
+# Or specify patterns and lower confidence threshold
+python generate_talib_annotations.py \
+  --input data/raw/OHLCV.csv \
+  --output talib_annotations.json \
+  --min-confidence 50 \
+  --patterns CDLENGULFING CDLHAMMER CDLDOJI
+
+# Exit container
+exit
+```
+
+This creates `talib_annotations.json` with detected patterns.
+
+### 1.3 Review Detection Results
+
+```bash
+# Check what was detected
+cat services/ml/talib_annotations.json | jq '.annotations | length'
+cat services/ml/talib_annotations.json | jq '.annotations[0]'
+
+# See pattern distribution
+cat services/ml/talib_annotations.json | jq '[.annotations[].label] | group_by(.) | map({label: .[0], count: length}) | sort_by(.count) | reverse'
+```
+
+**Output example:**
+```json
+{
+  "start_time": 1700000000,
+  "end_time": 1700003600,
+  "label": "Bullish Engulfing",
+  "confidence": 1.0,
+  "source": "programmatic",
+  "notes": "TA-Lib CDLENGULFING detection"
+}
+```
+
+## Step 2: Import into UI
+
+### 2.1 Copy Annotations File
+
+```bash
+# Copy from ML service to project root
+docker-compose cp ml-service:/app/talib_annotations.json ./talib_annotations.json
+```
+
+### 2.2 Get Your Chart ID
+
+Open http://localhost:3000 and note your chart ID (shown in the chart selector, or check database):
+
+```bash
+docker-compose exec candle-annotator sh -c "sqlite3 /app/data/candles.db 'SELECT id, name FROM charts;'"
+```
+
+### 2.3 Import Annotations
+
+```bash
+# Import into chart 1
+npm run import-annotations -- --file talib_annotations.json --chart-id 1
+
+# Or clear existing annotations first
+npm run import-annotations -- --file talib_annotations.json --chart-id 1 --clear
+```
+
+**Output:**
+```
+=== TA-Lib Annotation Import ===
+
+Input file: talib_annotations.json
+Chart ID: 1
+Clear existing: no
+
+Reading annotations file...
+Found 147 annotations
+Source: talib
+
+Ensuring 12 label types exist...
+  ✓ Bullish Engulfing (existing, id: 1)
+  + Bearish Engulfing (created, id: 8, color: #ef4444)
+  ✓ Bullish Hammer (existing, id: 2)
+  ...
+
+Importing 147 annotations for chart 1...
+
+✓ Imported 147 annotations
+
+=== Import Complete ===
+```
+
+## Step 3: Review & Edit in UI
+
+### 3.1 Open the Annotator
+
+1. Open http://localhost:3000
+2. Select your chart from the dropdown
+3. Scroll down to see the span annotations in the sidebar
+
+### 3.2 Review TA-Lib Detections
+
+You'll see all the TA-Lib detected patterns as span annotations:
+- Green spans = Bullish patterns
+- Red spans = Bearish patterns
+- Source labeled as "programmatic"
+
+### 3.3 Edit Annotations
+
+**Correct false positives:**
+- Click on a span annotation in the sidebar or chart
+- Press Delete/Backspace to remove it
+
+**Add missing patterns:**
+- Select a span label type
+- Click and drag on the chart to create new annotations
+- Your annotations are marked as source "human"
+
+**Adjust boundaries:**
+- Delete the annotation
+- Recreate it with correct start/end times
+
+**Add new pattern types:**
+- Go to "Manage Span Label Types"
+- Add custom patterns TA-Lib doesn't detect
+- Return to main page and annotate
+
+### 3.4 Best Practices
+
+- **Review all TA-Lib detections** - They're not always perfect
+- **Focus on quality over quantity** - Better to have 50 accurate annotations than 200 noisy ones
+- **Add context** - TA-Lib only detects classic patterns; add your own insights
+- **Diverse examples** - Make sure you have patterns in different market conditions
+
+## Step 4: Export & Train
+
+### 4.1 Export Annotations
+
+```bash
+# Export all span annotations (includes both human and TA-Lib)
+curl http://localhost:3000/api/span-annotations/export > services/ml/data/annotations/export.json
+
+# Verify export
+cat services/ml/data/annotations/export.json | jq '.annotations | length'
+```
+
+### 4.2 Prepare OHLCV Data
+
+```bash
+# Copy candles to ML service
+docker-compose exec candle-annotator sh -c "sqlite3 /app/data/candles.db -csv -header 'SELECT time, open, high, low, close, volume FROM candles ORDER BY time;'" > services/ml/data/raw/OHLCV.csv
+```
+
+### 4.3 Train Model
+
+```bash
+# Enter ML service
+docker-compose exec ml-service bash
+
+# Run full pipeline
+python pipeline.py --config config/pipeline.yaml
+
+# Exit
+exit
+```
+
+### 4.4 Restart Inference
+
+```bash
+# Restart to load new model
+docker-compose restart ml-service
+
+# Verify model loaded
+curl http://localhost:8001/model/info | jq '.model_info'
+```
+
+## Step 5: Compare & Iterate
+
+### 5.1 Get Predictions
+
+1. Open http://localhost:3000
+2. Scroll to Predictions panel
+3. Click "Run on Visible" or "Predict All"
+
+### 5.2 Compare with TA-Lib
+
+Now you can compare:
+- **Your edited annotations** (human judgment)
+- **TA-Lib raw detections** (programmatic)
+- **Model predictions** (trained on your corrections)
+
+The disagreement detection shows where these differ!
+
+### 5.3 Iterate
+
+Use the prediction summary to find:
+- **Missed by model** - Patterns you annotated but model missed
+- **Missed by human** - Model found patterns you didn't annotate
+- **Label mismatch** - Same location, different pattern type
+
+Add more annotations where the model struggles, then retrain.
+
+## Configuration Options
+
+### Pattern Selection
+
+Edit which patterns to detect:
+
+```bash
+python generate_talib_annotations.py \
+  --input data/raw/OHLCV.csv \
+  --output talib_annotations.json \
+  --patterns CDLENGULFING CDLHAMMER CDLDOJI CDLMORNINGSTAR CDLEVENINGSTAR
+```
+
+Common patterns:
+- `CDLENGULFING` - Bullish/Bearish Engulfing
+- `CDLHAMMER` - Hammer
+- `CDLDOJI` - Doji
+- `CDLMORNINGSTAR` / `CDLEVENINGSTAR` - Morning/Evening Star
+- `CDLHARAMI` - Harami
+- `CDLTHREEWHITESOLDIERS` / `CDLTHREEBLACKCROWS` - Three Soldiers/Crows
+
+See full list: https://ta-lib.org/function.html (search for CDL)
+
+### Confidence Threshold
+
+TA-Lib returns -100/+100 for pattern matches. Lower the threshold to get more detections:
+
+```bash
+# Get 50-100% matches (more patterns, potentially noisier)
+python generate_talib_annotations.py \
+  --input data/raw/OHLCV.csv \
+  --output talib_annotations.json \
+  --min-confidence 50
+```
+
+## Troubleshooting
+
+### "No module named 'talib'"
+
+TA-Lib not installed:
+
+```bash
+# Rebuild ml-service with TA-Lib
+docker-compose build --no-cache ml-service
+docker-compose up -d ml-service
+```
+
+### "No patterns detected"
+
+Try:
+1. **Lower confidence threshold** - Use `--min-confidence 50`
+2. **Check data quality** - Make sure OHLCV has valid data
+3. **Try more patterns** - Don't specify `--patterns`, detect all
+
+### Import script fails
+
+Make sure:
+1. **File exists** - Check path to JSON file
+2. **Chart ID valid** - Run: `docker-compose exec candle-annotator sh -c "sqlite3 /app/data/candles.db 'SELECT id, name FROM charts;'"`
+3. **tsx installed** - Run: `npm install`
+
+### Annotations not showing in UI
+
+1. **Refresh page** - Hard refresh (Ctrl+F5)
+2. **Check chart ID** - Make sure you selected the correct chart
+3. **Check database** - Run: `docker-compose exec candle-annotator sh -c "sqlite3 /app/data/candles.db 'SELECT COUNT(*) FROM span_annotations;'"`
+
+## Tips & Best Practices
+
+### Balancing Human & Programmatic Labels
+
+When training, you can choose merge strategy in `config/pipeline.yaml`:
+
+```yaml
+annotation_ingestion:
+  merge_strategy: "human_priority"  # Use human labels where they overlap
+  # or "programmatic_priority"      # Use TA-Lib where they overlap
+  # or "both"                       # Keep both as separate features
+```
+
+**Recommended**: Start with `human_priority` - trust your corrections over TA-Lib.
+
+### Iterative Improvement
+
+1. **Round 1**: Generate TA-Lib → Review → Train baseline model
+2. **Round 2**: Get predictions → Find disagreements → Add corrections → Retrain
+3. **Round 3**: Focus on low-confidence predictions → Add more examples → Retrain
+4. **Repeat** until model performance meets your needs
+
+### Pattern Coverage
+
+Make sure you have examples of:
+- **Bullish patterns** in uptrends
+- **Bearish patterns** in downtrends
+- **Neutral patterns** in sideways markets
+- **False signals** (TA-Lib detected but actually not a tradeable pattern)
+
+This teaches the model context, not just shape recognition.
+
+### Quality Metrics
+
+Track these in MLflow UI (http://localhost:5000):
+- **Accuracy** - Overall correctness
+- **F1 (macro)** - Average across all pattern types
+- **Per-class F1** - Performance for each pattern individually
+- **Confusion matrix** - Where the model makes mistakes
+
+Focus on improving F1 for patterns you actually trade.
+
+## Quick Reference
+
+```bash
+# Generate TA-Lib annotations
+docker-compose exec ml-service python generate_talib_annotations.py \
+  --input data/raw/OHLCV.csv --output talib_annotations.json
+
+# Copy to host
+docker-compose cp ml-service:/app/talib_annotations.json ./
+
+# Import to UI
+npm run import-annotations -- --file talib_annotations.json --chart-id 1
+
+# Export after editing
+curl http://localhost:3000/api/span-annotations/export > services/ml/data/annotations/export.json
+
+# Train model
+docker-compose exec ml-service python pipeline.py --config config/pipeline.yaml
+
+# Restart inference
+docker-compose restart ml-service
+
+# View results
+open http://localhost:3000
+open http://localhost:5000  # MLflow UI
+```
--- a/next-env.d.ts
+++ b/next-env.d.ts
@ -1,6 +1,6 @@
 /// <reference types="next" />
 /// <reference types="next/image-types/global" />
-import "./.next/types/routes.d.ts";
+import "./.next/dev/types/routes.d.ts";

 // NOTE: This file should not be edited
 // see https://nextjs.org/docs/app/api-reference/config/typescript for more information.
--- a/package-lock.json
+++ b/package-lock.json
@ -38,7 +38,8 @@
      },
      "devDependencies": {
        "drizzle-kit": "^0.31.9",
-        "tailwindcss": "^3.4.19"
+        "tailwindcss": "^3.4.19",
+        "tsx": "^4.21.0"
      }
    },
    "node_modules/@alloc/quick-lru": {
@ -7473,6 +7474,511 @@
      "version": "2.8.1",
      "license": "0BSD"
    },
+    "node_modules/tsx": {
+      "version": "4.21.0",
+      "resolved": "https://registry.npmjs.org/tsx/-/tsx-4.21.0.tgz",
+      "integrity": "sha512-5C1sg4USs1lfG0GFb2RLXsdpXqBSEhAaA/0kPL01wxzpMqLILNxIxIOKiILz+cdg/pLnOUxFYOR5yhHU666wbw==",
+      "dev": true,
+      "license": "MIT",
+      "peer": true,
+      "dependencies": {
+        "esbuild": "~0.27.0",
+        "get-tsconfig": "^4.7.5"
+      },
+      "bin": {
+        "tsx": "dist/cli.mjs"
+      },
+      "engines": {
+        "node": ">=18.0.0"
+      },
+      "optionalDependencies": {
+        "fsevents": "~2.3.3"
+      }
+    },
+    "node_modules/tsx/node_modules/@esbuild/aix-ppc64": {
+      "version": "0.27.3",
+      "resolved": "https://registry.npmjs.org/@esbuild/aix-ppc64/-/aix-ppc64-0.27.3.tgz",
+      "integrity": "sha512-9fJMTNFTWZMh5qwrBItuziu834eOCUcEqymSH7pY+zoMVEZg3gcPuBNxH1EvfVYe9h0x/Ptw8KBzv7qxb7l8dg==",
+      "cpu": [
+        "ppc64"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "aix"
+      ],
+      "engines": {
+        "node": ">=18"
+      }
+    },
+    "node_modules/tsx/node_modules/@esbuild/android-arm": {
+      "version": "0.27.3",
+      "resolved": "https://registry.npmjs.org/@esbuild/android-arm/-/android-arm-0.27.3.tgz",
+      "integrity": "sha512-i5D1hPY7GIQmXlXhs2w8AWHhenb00+GxjxRncS2ZM7YNVGNfaMxgzSGuO8o8SJzRc/oZwU2bcScvVERk03QhzA==",
+      "cpu": [
+        "arm"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "android"
+      ],
+      "engines": {
+        "node": ">=18"
+      }
+    },
+    "node_modules/tsx/node_modules/@esbuild/android-arm64": {
+      "version": "0.27.3",
+      "resolved": "https://registry.npmjs.org/@esbuild/android-arm64/-/android-arm64-0.27.3.tgz",
+      "integrity": "sha512-YdghPYUmj/FX2SYKJ0OZxf+iaKgMsKHVPF1MAq/P8WirnSpCStzKJFjOjzsW0QQ7oIAiccHdcqjbHmJxRb/dmg==",
+      "cpu": [
+        "arm64"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "android"
+      ],
+      "engines": {
+        "node": ">=18"
+      }
+    },
+    "node_modules/tsx/node_modules/@esbuild/android-x64": {
+      "version": "0.27.3",
+      "resolved": "https://registry.npmjs.org/@esbuild/android-x64/-/android-x64-0.27.3.tgz",
+      "integrity": "sha512-IN/0BNTkHtk8lkOM8JWAYFg4ORxBkZQf9zXiEOfERX/CzxW3Vg1ewAhU7QSWQpVIzTW+b8Xy+lGzdYXV6UZObQ==",
+      "cpu": [
+        "x64"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "android"
+      ],
+      "engines": {
+        "node": ">=18"
+      }
+    },
+    "node_modules/tsx/node_modules/@esbuild/darwin-arm64": {
+      "version": "0.27.3",
+      "resolved": "https://registry.npmjs.org/@esbuild/darwin-arm64/-/darwin-arm64-0.27.3.tgz",
+      "integrity": "sha512-Re491k7ByTVRy0t3EKWajdLIr0gz2kKKfzafkth4Q8A5n1xTHrkqZgLLjFEHVD+AXdUGgQMq+Godfq45mGpCKg==",
+      "cpu": [
+        "arm64"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "darwin"
+      ],
+      "engines": {
+        "node": ">=18"
+      }
+    },
+    "node_modules/tsx/node_modules/@esbuild/darwin-x64": {
+      "version": "0.27.3",
+      "resolved": "https://registry.npmjs.org/@esbuild/darwin-x64/-/darwin-x64-0.27.3.tgz",
+      "integrity": "sha512-vHk/hA7/1AckjGzRqi6wbo+jaShzRowYip6rt6q7VYEDX4LEy1pZfDpdxCBnGtl+A5zq8iXDcyuxwtv3hNtHFg==",
+      "cpu": [
+        "x64"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "darwin"
+      ],
+      "engines": {
+        "node": ">=18"
+      }
+    },
+    "node_modules/tsx/node_modules/@esbuild/freebsd-arm64": {
+      "version": "0.27.3",
+      "resolved": "https://registry.npmjs.org/@esbuild/freebsd-arm64/-/freebsd-arm64-0.27.3.tgz",
+      "integrity": "sha512-ipTYM2fjt3kQAYOvo6vcxJx3nBYAzPjgTCk7QEgZG8AUO3ydUhvelmhrbOheMnGOlaSFUoHXB6un+A7q4ygY9w==",
+      "cpu": [
+        "arm64"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "freebsd"
+      ],
+      "engines": {
+        "node": ">=18"
+      }
+    },
+    "node_modules/tsx/node_modules/@esbuild/freebsd-x64": {
+      "version": "0.27.3",
+      "resolved": "https://registry.npmjs.org/@esbuild/freebsd-x64/-/freebsd-x64-0.27.3.tgz",
+      "integrity": "sha512-dDk0X87T7mI6U3K9VjWtHOXqwAMJBNN2r7bejDsc+j03SEjtD9HrOl8gVFByeM0aJksoUuUVU9TBaZa2rgj0oA==",
+      "cpu": [
+        "x64"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "freebsd"
+      ],
+      "engines": {
+        "node": ">=18"
+      }
+    },
+    "node_modules/tsx/node_modules/@esbuild/linux-arm": {
+      "version": "0.27.3",
+      "resolved": "https://registry.npmjs.org/@esbuild/linux-arm/-/linux-arm-0.27.3.tgz",
+      "integrity": "sha512-s6nPv2QkSupJwLYyfS+gwdirm0ukyTFNl3KTgZEAiJDd+iHZcbTPPcWCcRYH+WlNbwChgH2QkE9NSlNrMT8Gfw==",
+      "cpu": [
+        "arm"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "linux"
+      ],
+      "engines": {
+        "node": ">=18"
+      }
+    },
+    "node_modules/tsx/node_modules/@esbuild/linux-arm64": {
+      "version": "0.27.3",
+      "resolved": "https://registry.npmjs.org/@esbuild/linux-arm64/-/linux-arm64-0.27.3.tgz",
+      "integrity": "sha512-sZOuFz/xWnZ4KH3YfFrKCf1WyPZHakVzTiqji3WDc0BCl2kBwiJLCXpzLzUBLgmp4veFZdvN5ChW4Eq/8Fc2Fg==",
+      "cpu": [
+        "arm64"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "linux"
+      ],
+      "engines": {
+        "node": ">=18"
+      }
+    },
+    "node_modules/tsx/node_modules/@esbuild/linux-ia32": {
+      "version": "0.27.3",
+      "resolved": "https://registry.npmjs.org/@esbuild/linux-ia32/-/linux-ia32-0.27.3.tgz",
+      "integrity": "sha512-yGlQYjdxtLdh0a3jHjuwOrxQjOZYD/C9PfdbgJJF3TIZWnm/tMd/RcNiLngiu4iwcBAOezdnSLAwQDPqTmtTYg==",
+      "cpu": [
+        "ia32"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "linux"
+      ],
+      "engines": {
+        "node": ">=18"
+      }
+    },
+    "node_modules/tsx/node_modules/@esbuild/linux-loong64": {
+      "version": "0.27.3",
+      "resolved": "https://registry.npmjs.org/@esbuild/linux-loong64/-/linux-loong64-0.27.3.tgz",
+      "integrity": "sha512-WO60Sn8ly3gtzhyjATDgieJNet/KqsDlX5nRC5Y3oTFcS1l0KWba+SEa9Ja1GfDqSF1z6hif/SkpQJbL63cgOA==",
+      "cpu": [
+        "loong64"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "linux"
+      ],
+      "engines": {
+        "node": ">=18"
+      }
+    },
+    "node_modules/tsx/node_modules/@esbuild/linux-mips64el": {
+      "version": "0.27.3",
+      "resolved": "https://registry.npmjs.org/@esbuild/linux-mips64el/-/linux-mips64el-0.27.3.tgz",
+      "integrity": "sha512-APsymYA6sGcZ4pD6k+UxbDjOFSvPWyZhjaiPyl/f79xKxwTnrn5QUnXR5prvetuaSMsb4jgeHewIDCIWljrSxw==",
+      "cpu": [
+        "mips64el"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "linux"
+      ],
+      "engines": {
+        "node": ">=18"
+      }
+    },
+    "node_modules/tsx/node_modules/@esbuild/linux-ppc64": {
+      "version": "0.27.3",
+      "resolved": "https://registry.npmjs.org/@esbuild/linux-ppc64/-/linux-ppc64-0.27.3.tgz",
+      "integrity": "sha512-eizBnTeBefojtDb9nSh4vvVQ3V9Qf9Df01PfawPcRzJH4gFSgrObw+LveUyDoKU3kxi5+9RJTCWlj4FjYXVPEA==",
+      "cpu": [
+        "ppc64"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "linux"
+      ],
+      "engines": {
+        "node": ">=18"
+      }
+    },
+    "node_modules/tsx/node_modules/@esbuild/linux-riscv64": {
+      "version": "0.27.3",
+      "resolved": "https://registry.npmjs.org/@esbuild/linux-riscv64/-/linux-riscv64-0.27.3.tgz",
+      "integrity": "sha512-3Emwh0r5wmfm3ssTWRQSyVhbOHvqegUDRd0WhmXKX2mkHJe1SFCMJhagUleMq+Uci34wLSipf8Lagt4LlpRFWQ==",
+      "cpu": [
+        "riscv64"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "linux"
+      ],
+      "engines": {
+        "node": ">=18"
+      }
+    },
+    "node_modules/tsx/node_modules/@esbuild/linux-s390x": {
+      "version": "0.27.3",
+      "resolved": "https://registry.npmjs.org/@esbuild/linux-s390x/-/linux-s390x-0.27.3.tgz",
+      "integrity": "sha512-pBHUx9LzXWBc7MFIEEL0yD/ZVtNgLytvx60gES28GcWMqil8ElCYR4kvbV2BDqsHOvVDRrOxGySBM9Fcv744hw==",
+      "cpu": [
+        "s390x"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "linux"
+      ],
+      "engines": {
+        "node": ">=18"
+      }
+    },
+    "node_modules/tsx/node_modules/@esbuild/linux-x64": {
+      "version": "0.27.3",
+      "resolved": "https://registry.npmjs.org/@esbuild/linux-x64/-/linux-x64-0.27.3.tgz",
+      "integrity": "sha512-Czi8yzXUWIQYAtL/2y6vogER8pvcsOsk5cpwL4Gk5nJqH5UZiVByIY8Eorm5R13gq+DQKYg0+JyQoytLQas4dA==",
+      "cpu": [
+        "x64"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "linux"
+      ],
+      "engines": {
+        "node": ">=18"
+      }
+    },
+    "node_modules/tsx/node_modules/@esbuild/netbsd-arm64": {
+      "version": "0.27.3",
+      "resolved": "https://registry.npmjs.org/@esbuild/netbsd-arm64/-/netbsd-arm64-0.27.3.tgz",
+      "integrity": "sha512-sDpk0RgmTCR/5HguIZa9n9u+HVKf40fbEUt+iTzSnCaGvY9kFP0YKBWZtJaraonFnqef5SlJ8/TiPAxzyS+UoA==",
+      "cpu": [
+        "arm64"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "netbsd"
+      ],
+      "engines": {
+        "node": ">=18"
+      }
+    },
+    "node_modules/tsx/node_modules/@esbuild/netbsd-x64": {
+      "version": "0.27.3",
+      "resolved": "https://registry.npmjs.org/@esbuild/netbsd-x64/-/netbsd-x64-0.27.3.tgz",
+      "integrity": "sha512-P14lFKJl/DdaE00LItAukUdZO5iqNH7+PjoBm+fLQjtxfcfFE20Xf5CrLsmZdq5LFFZzb5JMZ9grUwvtVYzjiA==",
+      "cpu": [
+        "x64"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "netbsd"
+      ],
+      "engines": {
+        "node": ">=18"
+      }
+    },
+    "node_modules/tsx/node_modules/@esbuild/openbsd-arm64": {
+      "version": "0.27.3",
+      "resolved": "https://registry.npmjs.org/@esbuild/openbsd-arm64/-/openbsd-arm64-0.27.3.tgz",
+      "integrity": "sha512-AIcMP77AvirGbRl/UZFTq5hjXK+2wC7qFRGoHSDrZ5v5b8DK/GYpXW3CPRL53NkvDqb9D+alBiC/dV0Fb7eJcw==",
+      "cpu": [
+        "arm64"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "openbsd"
+      ],
+      "engines": {
+        "node": ">=18"
+      }
+    },
+    "node_modules/tsx/node_modules/@esbuild/openbsd-x64": {
+      "version": "0.27.3",
+      "resolved": "https://registry.npmjs.org/@esbuild/openbsd-x64/-/openbsd-x64-0.27.3.tgz",
+      "integrity": "sha512-DnW2sRrBzA+YnE70LKqnM3P+z8vehfJWHXECbwBmH/CU51z6FiqTQTHFenPlHmo3a8UgpLyH3PT+87OViOh1AQ==",
+      "cpu": [
+        "x64"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "openbsd"
+      ],
+      "engines": {
+        "node": ">=18"
+      }
+    },
+    "node_modules/tsx/node_modules/@esbuild/openharmony-arm64": {
+      "version": "0.27.3",
+      "resolved": "https://registry.npmjs.org/@esbuild/openharmony-arm64/-/openharmony-arm64-0.27.3.tgz",
+      "integrity": "sha512-NinAEgr/etERPTsZJ7aEZQvvg/A6IsZG/LgZy+81wON2huV7SrK3e63dU0XhyZP4RKGyTm7aOgmQk0bGp0fy2g==",
+      "cpu": [
+        "arm64"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "openharmony"
+      ],
+      "engines": {
+        "node": ">=18"
+      }
+    },
+    "node_modules/tsx/node_modules/@esbuild/sunos-x64": {
+      "version": "0.27.3",
+      "resolved": "https://registry.npmjs.org/@esbuild/sunos-x64/-/sunos-x64-0.27.3.tgz",
+      "integrity": "sha512-PanZ+nEz+eWoBJ8/f8HKxTTD172SKwdXebZ0ndd953gt1HRBbhMsaNqjTyYLGLPdoWHy4zLU7bDVJztF5f3BHA==",
+      "cpu": [
+        "x64"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "sunos"
+      ],
+      "engines": {
+        "node": ">=18"
+      }
+    },
+    "node_modules/tsx/node_modules/@esbuild/win32-arm64": {
+      "version": "0.27.3",
+      "resolved": "https://registry.npmjs.org/@esbuild/win32-arm64/-/win32-arm64-0.27.3.tgz",
+      "integrity": "sha512-B2t59lWWYrbRDw/tjiWOuzSsFh1Y/E95ofKz7rIVYSQkUYBjfSgf6oeYPNWHToFRr2zx52JKApIcAS/D5TUBnA==",
+      "cpu": [
+        "arm64"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "win32"
+      ],
+      "engines": {
+        "node": ">=18"
+      }
+    },
+    "node_modules/tsx/node_modules/@esbuild/win32-ia32": {
+      "version": "0.27.3",
+      "resolved": "https://registry.npmjs.org/@esbuild/win32-ia32/-/win32-ia32-0.27.3.tgz",
+      "integrity": "sha512-QLKSFeXNS8+tHW7tZpMtjlNb7HKau0QDpwm49u0vUp9y1WOF+PEzkU84y9GqYaAVW8aH8f3GcBck26jh54cX4Q==",
+      "cpu": [
+        "ia32"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "win32"
+      ],
+      "engines": {
+        "node": ">=18"
+      }
+    },
+    "node_modules/tsx/node_modules/@esbuild/win32-x64": {
+      "version": "0.27.3",
+      "resolved": "https://registry.npmjs.org/@esbuild/win32-x64/-/win32-x64-0.27.3.tgz",
+      "integrity": "sha512-4uJGhsxuptu3OcpVAzli+/gWusVGwZZHTlS63hh++ehExkVT8SgiEf7/uC/PclrPPkLhZqGgCTjd0VWLo6xMqA==",
+      "cpu": [
+        "x64"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "win32"
+      ],
+      "engines": {
+        "node": ">=18"
+      }
+    },
+    "node_modules/tsx/node_modules/esbuild": {
+      "version": "0.27.3",
+      "resolved": "https://registry.npmjs.org/esbuild/-/esbuild-0.27.3.tgz",
+      "integrity": "sha512-8VwMnyGCONIs6cWue2IdpHxHnAjzxnw2Zr7MkVxB2vjmQ2ivqGFb4LEG3SMnv0Gb2F/G/2yA8zUaiL1gywDCCg==",
+      "dev": true,
+      "hasInstallScript": true,
+      "license": "MIT",
+      "bin": {
+        "esbuild": "bin/esbuild"
+      },
+      "engines": {
+        "node": ">=18"
+      },
+      "optionalDependencies": {
+        "@esbuild/aix-ppc64": "0.27.3",
+        "@esbuild/android-arm": "0.27.3",
+        "@esbuild/android-arm64": "0.27.3",
+        "@esbuild/android-x64": "0.27.3",
+        "@esbuild/darwin-arm64": "0.27.3",
+        "@esbuild/darwin-x64": "0.27.3",
+        "@esbuild/freebsd-arm64": "0.27.3",
+        "@esbuild/freebsd-x64": "0.27.3",
+        "@esbuild/linux-arm": "0.27.3",
+        "@esbuild/linux-arm64": "0.27.3",
+        "@esbuild/linux-ia32": "0.27.3",
+        "@esbuild/linux-loong64": "0.27.3",
+        "@esbuild/linux-mips64el": "0.27.3",
+        "@esbuild/linux-ppc64": "0.27.3",
+        "@esbuild/linux-riscv64": "0.27.3",
+        "@esbuild/linux-s390x": "0.27.3",
+        "@esbuild/linux-x64": "0.27.3",
+        "@esbuild/netbsd-arm64": "0.27.3",
+        "@esbuild/netbsd-x64": "0.27.3",
+        "@esbuild/openbsd-arm64": "0.27.3",
+        "@esbuild/openbsd-x64": "0.27.3",
+        "@esbuild/openharmony-arm64": "0.27.3",
+        "@esbuild/sunos-x64": "0.27.3",
+        "@esbuild/win32-arm64": "0.27.3",
+        "@esbuild/win32-ia32": "0.27.3",
+        "@esbuild/win32-x64": "0.27.3"
+      }
+    },
    "node_modules/tunnel-agent": {
      "version": "0.6.0",
      "license": "Apache-2.0",
--- a/package.json
+++ b/package.json
@ -7,7 +7,8 @@
    "dev": "next dev",
    "build": "next build",
    "start": "next start",
-    "lint": "next lint"
+    "lint": "next lint",
+    "import-annotations": "tsx scripts/import_talib_annotations.ts"
  },
  "keywords": [],
  "author": "",
@ -42,6 +43,7 @@
  },
  "devDependencies": {
    "drizzle-kit": "^0.31.9",
-    "tailwindcss": "^3.4.19"
+    "tailwindcss": "^3.4.19",
+    "tsx": "^4.21.0"
  }
 }
--- a/scripts/import_talib_annotations.ts
+++ b/scripts/import_talib_annotations.ts
@ -0,0 +1,252 @@
+#!/usr/bin/env tsx
+/**
+ * Import TA-Lib generated annotations into the Candle Annotator database.
+ * 
+ * This script reads a JSON file with annotations (from generate_talib_annotations.py)
+ * and imports them as span annotations that can be viewed and edited in the UI.
+ * 
+ * Usage:
+ *   npm run import-annotations -- --file talib_annotations.json --chart-id 1
+ */
+
+import { readFile } from 'fs/promises';
+import { db } from '../src/lib/db';
+import { spanAnnotations, spanLabelTypes } from '../src/lib/db/schema';
+import { eq } from 'drizzle-orm';
+
+interface Annotation {
+  start_time: number;
+  end_time: number;
+  label: string;
+  confidence?: number;
+  source?: string;
+  notes?: string;
+}
+
+interface ImportData {
+  annotations: Annotation[];
+  metadata?: {
+    source?: string;
+    count?: number;
+  };
+}
+
+interface LabelTypeMap {
+  [key: string]: {
+    id: number;
+    color: string;
+  };
+}
+
+async function ensureLabelTypes(labels: string[]): Promise<LabelTypeMap> {
+  /**
+   * Ensure span label types exist for all unique labels.
+   * Creates missing label types with auto-generated colors.
+   */
+  const uniqueLabels = [...new Set(labels)];
+  const labelMap: LabelTypeMap = {};
+  
+  console.log(`\nEnsuring ${uniqueLabels.length} label types exist...`);
+  
+  // Fetch existing label types
+  const existing = await db.select().from(spanLabelTypes);
+  const existingMap = new Map(existing.map(lt => [lt.name, lt]));
+  
+  // Color palette for auto-generated labels
+  const colors = [
+    '#22c55e', // green (bullish)
+    '#ef4444', // red (bearish)
+    '#3b82f6', // blue
+    '#f59e0b', // amber
+    '#8b5cf6', // purple
+    '#ec4899', // pink
+    '#06b6d4', // cyan
+    '#f97316', // orange
+  ];
+  
+  let colorIndex = 0;
+  let sortOrder = existing.length;
+  
+  for (const label of uniqueLabels) {
+    if (existingMap.has(label)) {
+      const existing = existingMap.get(label)!;
+      labelMap[label] = {
+        id: existing.id,
+        color: existing.color,
+      };
+      console.log(`  ✓ ${label} (existing, id: ${existing.id})`);
+    } else {
+      // Assign color based on bullish/bearish
+      let color: string;
+      if (label.toLowerCase().includes('bullish')) {
+        color = '#22c55e'; // green
+      } else if (label.toLowerCase().includes('bearish')) {
+        color = '#ef4444'; // red
+      } else {
+        color = colors[colorIndex % colors.length];
+        colorIndex++;
+      }
+      
+      // Create new label type
+      const [newLabel] = await db.insert(spanLabelTypes).values({
+        name: label,
+        display_name: label,
+        color,
+        hotkey: null,
+        is_active: 1,
+        sort_order: sortOrder++,
+      }).returning();
+      
+      labelMap[label] = {
+        id: newLabel.id,
+        color: newLabel.color,
+      };
+      
+      console.log(`  + ${label} (created, id: ${newLabel.id}, color: ${color})`);
+    }
+  }
+  
+  return labelMap;
+}
+
+async function importAnnotations(
+  annotations: Annotation[],
+  chartId: number,
+  labelMap: LabelTypeMap
+): Promise<number> {
+  /**
+   * Import annotations into the database.
+   * Returns the number of annotations imported.
+   */
+  console.log(`\nImporting ${annotations.length} annotations for chart ${chartId}...`);
+  
+  let imported = 0;
+  let skipped = 0;
+  
+  for (const ann of annotations) {
+    const labelInfo = labelMap[ann.label];
+    if (!labelInfo) {
+      console.error(`  ✗ Skipping: Unknown label "${ann.label}"`);
+      skipped++;
+      continue;
+    }
+    
+    try {
+      await db.insert(spanAnnotations).values({
+        chart_id: chartId,
+        start_time: ann.start_time,
+        end_time: ann.end_time,
+        label: ann.label,
+        confidence: ann.confidence || null,
+        outcome: null,
+        notes: ann.notes || null,
+        sub_spans: null,
+        color: labelInfo.color,
+        source: ann.source || 'programmatic',
+        model_prediction: null,
+      });
+      
+      imported++;
+      
+      if (imported % 100 === 0) {
+        console.log(`  ${imported}/${annotations.length} imported...`);
+      }
+    } catch (error: any) {
+      console.error(`  ✗ Error importing annotation: ${error.message}`);
+      skipped++;
+    }
+  }
+  
+  console.log(`\n✓ Imported ${imported} annotations`);
+  if (skipped > 0) {
+    console.log(`✗ Skipped ${skipped} annotations`);
+  }
+  
+  return imported;
+}
+
+async function main() {
+  const args = process.argv.slice(2);
+  
+  // Parse arguments
+  let filePath: string | null = null;
+  let chartId: number | null = null;
+  let clearExisting = false;
+  
+  for (let i = 0; i < args.length; i++) {
+    if (args[i] === '--file' || args[i] === '-f') {
+      filePath = args[++i];
+    } else if (args[i] === '--chart-id' || args[i] === '-c') {
+      chartId = parseInt(args[++i], 10);
+    } else if (args[i] === '--clear') {
+      clearExisting = true;
+    } else if (args[i] === '--help' || args[i] === '-h') {
+      console.log(`
+Import TA-Lib annotations into Candle Annotator database
+
+Usage:
+  npm run import-annotations -- --file <json-file> --chart-id <id> [--clear]
+
+Options:
+  --file, -f <path>     Input JSON file (from generate_talib_annotations.py)
+  --chart-id, -c <id>   Chart ID to import annotations into
+  --clear               Clear existing annotations for this chart before import
+  --help, -h            Show this help message
+
+Example:
+  npm run import-annotations -- --file talib_annotations.json --chart-id 1
+      `);
+      process.exit(0);
+    }
+  }
+  
+  if (!filePath || chartId === null) {
+    console.error('Error: --file and --chart-id are required');
+    console.error('Run with --help for usage information');
+    process.exit(1);
+  }
+  
+  console.log('=== TA-Lib Annotation Import ===\n');
+  console.log(`Input file: ${filePath}`);
+  console.log(`Chart ID: ${chartId}`);
+  console.log(`Clear existing: ${clearExisting ? 'yes' : 'no'}`);
+  
+  // Read annotations file
+  console.log('\nReading annotations file...');
+  const fileContent = await readFile(filePath, 'utf-8');
+  const data: ImportData = JSON.parse(fileContent);
+  
+  console.log(`Found ${data.annotations.length} annotations`);
+  if (data.metadata) {
+    console.log(`Source: ${data.metadata.source || 'unknown'}`);
+  }
+  
+  // Clear existing annotations if requested
+  if (clearExisting) {
+    console.log('\nClearing existing annotations...');
+    const result = await db.delete(spanAnnotations)
+      .where(eq(spanAnnotations.chart_id, chartId));
+    console.log(`Deleted existing annotations`);
+  }
+  
+  // Collect unique labels
+  const uniqueLabels = [...new Set(data.annotations.map(a => a.label))];
+  
+  // Ensure label types exist
+  const labelMap = await ensureLabelTypes(uniqueLabels);
+  
+  // Import annotations
+  const imported = await importAnnotations(data.annotations, chartId, labelMap);
+  
+  console.log('\n=== Import Complete ===');
+  console.log(`\nNext steps:`);
+  console.log(`1. Open http://localhost:3000 and select chart ${chartId}`);
+  console.log(`2. Review the TA-Lib generated annotations`);
+  console.log(`3. Edit, delete, or add new annotations as needed`);
+  console.log(`4. Export and train: npm run ml:export-and-train`);
+}
+
+main().catch((error) => {
+  console.error('\n✗ Import failed:', error.message);
+  process.exit(1);
+});
--- a/services/ml/app/pycache/init.cpython-313.pyc
+++ b/services/ml/app/pycache/init.cpython-313.pyc
--- a/services/ml/app/pycache/config.cpython-313.pyc
+++ b/services/ml/app/pycache/config.cpython-313.pyc
--- a/services/ml/app/pycache/main.cpython-313.pyc
+++ b/services/ml/app/pycache/main.cpython-313.pyc
--- a/services/ml/app/pycache/preprocessing.cpython-313.pyc
+++ b/services/ml/app/pycache/preprocessing.cpython-313.pyc
--- a/services/ml/candle_ml.egg-info/SOURCES.txt
+++ b/services/ml/candle_ml.egg-info/SOURCES.txt
@ -1,6 +1,22 @@
 pyproject.toml
+app/__init__.py
+app/annotation_ingestion.py
+app/config.py
+app/db.py
+app/main.py
+app/preprocessing.py
 candle_ml.egg-info/PKG-INFO
 candle_ml.egg-info/SOURCES.txt
 candle_ml.egg-info/dependency_links.txt
 candle_ml.egg-info/requires.txt
-candle_ml.egg-info/top_level.txt
+candle_ml.egg-info/top_level.txt
+features/candle_features.py
+features/custom_loader.py
+features/engineer.py
+features/talib_features.py
+training/__init__.py
+training/evaluation.py
+training/train.py
+training/models/__init__.py
+training/models/random_forest.py
+training/models/xgboost_model.py
--- a/services/ml/features/pycache/candle_features.cpython-313.pyc
+++ b/services/ml/features/pycache/candle_features.cpython-313.pyc
--- a/services/ml/features/pycache/custom_loader.cpython-313.pyc
+++ b/services/ml/features/pycache/custom_loader.cpython-313.pyc
--- a/services/ml/features/pycache/talib_features.cpython-313.pyc
+++ b/services/ml/features/pycache/talib_features.cpython-313.pyc
--- a/services/ml/generate_talib_annotations.py
+++ b/services/ml/generate_talib_annotations.py
@ -0,0 +1,267 @@
+#!/usr/bin/env python3
+"""
+Generate span annotations from TA-Lib candlestick pattern functions.
+
+This script runs TA-Lib CDL* functions on OHLCV data and outputs
+annotations in a format that can be imported into the Candle Annotator UI.
+
+Usage:
+    python generate_talib_annotations.py --input data/raw/OHLCV.csv --output talib_annotations.json
+"""
+
+import argparse
+import json
+import logging
+from pathlib import Path
+from typing import List, Dict, Any
+
+import pandas as pd
+import numpy as np
+
+try:
+    import talib
+except ImportError:
+    print("ERROR: TA-Lib not installed. Install with: pip install TA-Lib")
+    print("Note: You may need to install the C library first. See DEPLOYMENT.md")
+    exit(1)
+
+logging.basicConfig(level=logging.INFO, format='[%(levelname)s] %(message)s')
+logger = logging.getLogger(__name__)
+
+# TA-Lib candlestick pattern functions with friendly names
+TALIB_PATTERNS = {
+    'CDLENGULFING': 'Engulfing',
+    'CDLHAMMER': 'Hammer',
+    'CDLINVERTEDHAMMER': 'Inverted Hammer',
+    'CDLSHOOTINGSTAR': 'Shooting Star',
+    'CDLDOJI': 'Doji',
+    'CDLDOJISTAR': 'Doji Star',
+    'CDLMORNINGSTAR': 'Morning Star',
+    'CDLEVENINGSTAR': 'Evening Star',
+    'CDLHARAMI': 'Harami',
+    'CDLHARAMICROSS': 'Harami Cross',
+    'CDLPIERCING': 'Piercing',
+    'CDLDARKCLOUDCOVER': 'Dark Cloud Cover',
+    'CDLTHREEWHITESOLDIERS': 'Three White Soldiers',
+    'CDLTHREEBLACKCROWS': 'Three Black Crows',
+    'CDLMARUBOZU': 'Marubozu',
+    'CDLSPINNINGTOP': 'Spinning Top',
+    'CDL3BLACKCROWS': 'Three Black Crows',
+    'CDL3WHITESOLDIERS': 'Three White Soldiers',
+    'CDLABANDONEDBABY': 'Abandoned Baby',
+    'CDLADVANCEBLOCK': 'Advance Block',
+    'CDLBELTHOLD': 'Belt Hold',
+    'CDLBREAKAWAY': 'Breakaway',
+    'CDLCLOSINGMARUBOZU': 'Closing Marubozu',
+    'CDLCONCEALBABYSWALL': 'Concealing Baby Swallow',
+    'CDLCOUNTERATTACK': 'Counterattack',
+    'CDLDRAGONFLYDOJI': 'Dragonfly Doji',
+    'CDLGAPSIDESIDEWHITE': 'Up/Down Gap Side-by-Side White Lines',
+    'CDLGRAVESTONEDOJI': 'Gravestone Doji',
+    'CDLHANGINGMAN': 'Hanging Man',
+    'CDLHIGHWAVE': 'High Wave',
+    'CDLHIKKAKE': 'Hikkake',
+    'CDLHIKKAKEMOD': 'Modified Hikkake',
+    'CDLHOMINGPIGEON': 'Homing Pigeon',
+    'CDLIDENTICAL3CROWS': 'Identical Three Crows',
+    'CDLINNECK': 'In-Neck',
+    'CDLKICKING': 'Kicking',
+    'CDLKICKINGBYLENGTH': 'Kicking by Length',
+    'CDLLADDERBOTTOM': 'Ladder Bottom',
+    'CDLLONGLEGGEDDOJI': 'Long-Legged Doji',
+    'CDLLONGLINE': 'Long Line',
+    'CDLMATCHINGLOW': 'Matching Low',
+    'CDLMATHOLD': 'Mat Hold',
+    'CDLMORNINGDOJISTAR': 'Morning Doji Star',
+    'CDLONNECK': 'On-Neck',
+    'CDLRISEFALL3METHODS': 'Rising/Falling Three Methods',
+    'CDLSEPARATINGLINES': 'Separating Lines',
+    'CDLSHORTLINE': 'Short Line',
+    'CDLSTALLEDPATTERN': 'Stalled Pattern',
+    'CDLSTICKSANDWICH': 'Stick Sandwich',
+    'CDLTAKURI': 'Takuri',
+    'CDLTASUKIGAP': 'Tasuki Gap',
+    'CDLTHRUSTING': 'Thrusting',
+    'CDLTRISTAR': 'Tristar',
+    'CDLUNIQUE3RIVER': 'Unique Three River',
+    'CDLUPSIDEGAP2CROWS': 'Upside Gap Two Crows',
+    'CDLXSIDEGAP3METHODS': 'Upside/Downside Gap Three Methods',
+}
+
+
+def load_ohlcv(input_path: str) -> pd.DataFrame:
+    """
+    Load OHLCV data from CSV file.
+    
+    Expected columns: time, open, high, low, close[, volume]
+    """
+    logger.info(f"Loading OHLCV data from {input_path}")
+    df = pd.read_csv(input_path)
+    
+    required_cols = ['time', 'open', 'high', 'low', 'close']
+    missing = [col for col in required_cols if col not in df.columns]
+    if missing:
+        raise ValueError(f"Missing required columns: {missing}")
+    
+    logger.info(f"Loaded {len(df)} candles")
+    return df
+
+
+def detect_talib_patterns(df: pd.DataFrame, patterns: List[str] = None) -> pd.DataFrame:
+    """
+    Run TA-Lib CDL* functions on OHLCV data.
+    
+    Args:
+        df: DataFrame with open, high, low, close columns
+        patterns: List of pattern names to detect (default: all)
+    
+    Returns:
+        DataFrame with pattern detection results
+    """
+    if patterns is None:
+        patterns = list(TALIB_PATTERNS.keys())
+    
+    results = df[['time', 'open', 'high', 'low', 'close']].copy()
+    
+    open_prices = df['open'].values
+    high_prices = df['high'].values
+    low_prices = df['low'].values
+    close_prices = df['close'].values
+    
+    logger.info(f"Running {len(patterns)} TA-Lib pattern detection functions...")
+    
+    for pattern_func in patterns:
+        if not hasattr(talib, pattern_func):
+            logger.warning(f"Unknown TA-Lib function: {pattern_func}")
+            continue
+        
+        try:
+            func = getattr(talib, pattern_func)
+            pattern_values = func(open_prices, high_prices, low_prices, close_prices)
+            results[pattern_func] = pattern_values
+        except Exception as e:
+            logger.error(f"Error running {pattern_func}: {e}")
+            results[pattern_func] = 0
+    
+    return results
+
+
+def create_span_annotations(df: pd.DataFrame, min_confidence: int = 100) -> List[Dict[str, Any]]:
+    """
+    Convert TA-Lib pattern detection results to span annotations.
+    
+    TA-Lib returns:
+    - 0: No pattern
+    - 100: Bullish pattern
+    - -100: Bearish pattern
+    
+    Args:
+        df: DataFrame with pattern detection columns
+        min_confidence: Minimum absolute confidence (100 = only perfect matches)
+    
+    Returns:
+        List of annotation dicts
+    """
+    annotations = []
+    
+    pattern_cols = [col for col in df.columns if col.startswith('CDL')]
+    
+    for idx, row in df.iterrows():
+        for pattern_col in pattern_cols:
+            pattern_value = row[pattern_col]
+            
+            # Skip if no pattern detected
+            if pattern_value == 0 or abs(pattern_value) < min_confidence:
+                continue
+            
+            # Determine label
+            friendly_name = TALIB_PATTERNS.get(pattern_col, pattern_col)
+            if pattern_value > 0:
+                label = f"Bullish {friendly_name}"
+            else:
+                label = f"Bearish {friendly_name}"
+            
+            # Create annotation
+            # TA-Lib patterns are single-candle or small multi-candle patterns
+            # We'll use a 3-candle span centered on the detection point
+            start_idx = max(0, idx - 1)
+            end_idx = min(len(df) - 1, idx + 1)
+            
+            start_time = int(df.iloc[start_idx]['time'])
+            end_time = int(df.iloc[end_idx]['time'])
+            
+            annotation = {
+                'start_time': start_time,
+                'end_time': end_time,
+                'label': label,
+                'confidence': abs(pattern_value) / 100.0,  # Normalize to 0-1
+                'source': 'programmatic',
+                'notes': f'TA-Lib {pattern_col} detection',
+            }
+            
+            annotations.append(annotation)
+    
+    return annotations
+
+
+def save_annotations(annotations: List[Dict[str, Any]], output_path: str):
+    """
+    Save annotations to JSON file in format compatible with the UI.
+    """
+    output_data = {
+        'annotations': annotations,
+        'metadata': {
+            'source': 'talib',
+            'count': len(annotations),
+        }
+    }
+    
+    output_file = Path(output_path)
+    output_file.parent.mkdir(parents=True, exist_ok=True)
+    
+    with open(output_file, 'w') as f:
+        json.dump(output_data, f, indent=2)
+    
+    logger.info(f"Saved {len(annotations)} annotations to {output_path}")
+
+
+def main():
+    parser = argparse.ArgumentParser(description='Generate span annotations from TA-Lib patterns')
+    parser.add_argument('--input', '-i', required=True, help='Input OHLCV CSV file')
+    parser.add_argument('--output', '-o', required=True, help='Output JSON file')
+    parser.add_argument('--min-confidence', type=int, default=100, 
+                        help='Minimum confidence (0-100, default: 100 = perfect matches only)')
+    parser.add_argument('--patterns', nargs='+', 
+                        help='Specific patterns to detect (default: all)')
+    
+    args = parser.parse_args()
+    
+    # Load data
+    df = load_ohlcv(args.input)
+    
+    # Detect patterns
+    patterns = args.patterns if args.patterns else list(TALIB_PATTERNS.keys())
+    results_df = detect_talib_patterns(df, patterns)
+    
+    # Create annotations
+    annotations = create_span_annotations(results_df, min_confidence=args.min_confidence)
+    
+    logger.info(f"Found {len(annotations)} pattern annotations")
+    
+    # Show label distribution
+    from collections import Counter
+    label_counts = Counter(ann['label'] for ann in annotations)
+    logger.info("Pattern distribution:")
+    for label, count in label_counts.most_common():
+        logger.info(f"  {label}: {count}")
+    
+    # Save
+    save_annotations(annotations, args.output)
+    
+    logger.info("\nNext steps:")
+    logger.info(f"1. Review/edit annotations: Use the import script (coming soon)")
+    logger.info(f"2. Or use directly for training: Copy to services/ml/data/annotations/export.json")
+
+
+if __name__ == '__main__':
+    main()
--- a/services/ml/mlruns/0/meta.yaml
+++ b/services/ml/mlruns/0/meta.yaml
@ -0,0 +1,6 @@
+artifact_location: /home/homoludens/projekti/bitcon/candle_annotator/services/ml/mlruns/artifacts/0
+creation_time: 1771178634585
+experiment_id: '0'
+last_update_time: 1771178634585
+lifecycle_stage: active
+name: Default
--- a/services/ml/uv.lock
+++ b/services/ml/uv.lock