generate_form_guides.py

"""Generate comprehensive form guide for all horses racing today."""
import json
from pathlib import Path
from collections import defaultdict
from datetime import datetime

DATA_DIR = Path('/home/brettanthonysjoberg179/trifecta-bro-hf-space/data')
CACHE_DIR = DATA_DIR / 'cache'
RESULTS_DIR = DATA_DIR / 'results'
REPORTS_DIR = DATA_DIR / 'reports'

VAULT_DIR = Path('/home/brettanthonysjoberg179/Documents/Obsidian Vault')
DATABANK_DIR = VAULT_DIR / 'Horse Form Databank'
FORM_GUIDES_DIR = DATABANK_DIR / 'Form-Guides'


def ensure_dirs():
    FORM_GUIDES_DIR.mkdir(parents=True, exist_ok=True)


def load_all_cached_races(date: str) -> list[dict]:
    races = []
    for f in CACHE_DIR.glob('*.json'):
        try:
            with open(f) as fp:
                data = json.load(fp)
            if data.get('date') == date and 'track' in data:
                races.append(data)
        except:
            continue
    return races


def parse_num(s):
    return int(str(s).replace('e', '').replace('E', ''))


def generate_horse_form_guide(horse: dict) -> str:
    """Generate comprehensive form guide for a single horse."""
    name = horse.get('name', 'Unknown')
    number = horse.get('number', 0)
    barrier = horse.get('barrier', 0)
    weight = horse.get('weight', 0)
    age = horse.get('age', 0)
    sex = horse.get('sex', '')
    sire = horse.get('sire', '')
    dam = horse.get('dam', '')
    jockey = horse.get('jockey', '')
    trainer = horse.get('trainer', '')
    form = horse.get('form', '')
    last20 = horse.get('last20_starts', '')
    career_prize = horse.get('career_prize_money', '$0')
    
    stats = horse.get('stats', {})
    overall = stats.get('overall', {})
    track = stats.get('track', {})
    distance = stats.get('distance', {})
    track_dist = stats.get('trackDistance', {})
    condition = stats.get('condition', {})
    first_up = stats.get('firstUp', {})
    second_up = stats.get('secondUp', {})
    
    # Get all condition stats
    cond_stats = stats.get('conditions', {})
    
    note = f"""# {name}

## Race Details

- **Number**: {number}
- **Barrier**: {barrier}
- **Weight**: {weight}kg
- **Jockey**: {jockey}
- **Trainer**: {trainer}
- **Age/Sex**: {age}yo {sex}
- **Sire**: {sire}
- **Dam**: {dam}
- **Career Prize**: {career_prize}

## Career Statistics

| Metric | Overall | Track | Distance | Track+Dist | Condition |
|--------|---------|-------|----------|------------|-----------|
| Starts | {overall.get('starts', 0)} | {track.get('starts', 0)} | {distance.get('starts', 0)} | {track_dist.get('starts', 0)} | {condition.get('starts', 0)} |
| Wins | {overall.get('wins', 0)} | {track.get('wins', 0)} | {distance.get('wins', 0)} | {track_dist.get('wins', 0)} | {condition.get('wins', 0)} |
| Places | {overall.get('places', 0)} | {track.get('places', 0)} | {distance.get('places', 0)} | {track_dist.get('places', 0)} | {condition.get('places', 0)} |
| Win % | {overall.get('winPercent', 0)*100:.1f}% | {track.get('winPercent', 0)*100:.1f}% | {distance.get('winPercent', 0)*100:.1f}% | {track_dist.get('winPercent', 0)*100:.1f}% | {condition.get('winPercent', 0)*100:.1f}% |
| Place % | {overall.get('placePercent', 0)*100:.1f}% | {track.get('placePercent', 0)*100:.1f}% | {distance.get('placePercent', 0)*100:.1f}% | {track_dist.get('placePercent', 0)*100:.1f}% | {condition.get('placePercent', 0)*100:.1f}% |

## Fitness Profile

| First-Up | Second-Up |
|----------|-----------|
| {first_up.get('starts', 0)} starts, {first_up.get('wins', 0)} wins ({first_up.get('winPercent', 0)*100:.1f}%) | {second_up.get('starts', 0)} starts, {second_up.get('wins', 0)} wins ({second_up.get('winPercent', 0)*100:.1f}%) |

## Condition Performance

| Condition | Starts | Wins | Win % | Place % |
|-----------|--------|------|-------|---------|
"""
    
    for cond_name, cond_data in sorted(cond_stats.items()):
        if cond_data.get('starts', 0) > 0:
            note += f"| {cond_name} | {cond_data.get('starts', 0)} | {cond_data.get('wins', 0)} | {cond_data.get('winPercent', 0)*100:.1f}% | {cond_data.get('placePercent', 0)*100:.1f}% |\n"
    
    note += f"""

## Last 20 Starts

```
{last20}
```

## Form Analysis

```
{form}
```

### Form Key
- **1-3** = Finished in top 3
- **4-9** = Finished 4th-9th
- **0** = Finished 10th or worse
- **x** = Spell (not raced recently)
- **f** = Fell
- **p** = Pulled up

---

## Notes

[Add your observations here]

## Related Horses

[Link to related horses: sire, dam, stablemates]
"""
    
    return note


def generate_race_form_guide(race_data: dict, horses: list[dict]) -> str:
    """Generate comprehensive form guide for a race."""
    date = race_data.get('date', '')
    track = race_data.get('track', '')
    race_num = race_data.get('raceNumber', 0)
    race_name = race_data.get('raceName', '')
    distance = race_data.get('distance', '')
    condition = race_data.get('condition', '')
    race_class = race_data.get('raceClass', '')
    prize = race_data.get('prizeMoney', '')
    start_time = race_data.get('startTime', '')
    num_runners = race_data.get('numberOfRunners', 0)
    
    note = f"""# {track} Race {race_num} — {date}

## Race Details

- **Name**: {race_name}
- **Distance**: {distance}
- **Condition**: {condition}
- **Class**: {race_class}
- **Prize**: {prize}
- **Start**: {start_time}
- **Runners**: {num_runners}

## Form Guide

| # | Horse | Jockey | Wgt | Barrier | Form | Win% | Place% | Track Win% | Dist Win% |
|---|-------|--------|-----|---------|------|------|--------|------------|-----------|
"""
    
    for horse in horses:
        stats = horse.get('stats', {})
        overall = stats.get('overall', {})
        track = stats.get('track', {})
        distance = stats.get('distance', {})
        
        note += f"| {horse['number']} | [[{horse['name']}]] | {horse.get('jockey', '')} | {horse.get('weight', 0)}kg | {horse.get('barrier', 0)} | {horse.get('form', '')} | {overall.get('winPercent', 0)*100:.1f}% | {overall.get('placePercent', 0)*100:.1f}% | {track.get('winPercent', 0)*100:.1f}% | {distance.get('winPercent', 0)*100:.1f}% |\n"
    
    note += """

## Result

[Add official result here]

## Analysis

[Add race analysis here]
"""
    
    return note


def main():
    date = "2026-08-15"
    print(f"Generating comprehensive form guides for {date}...")
    print("=" * 60)
    
    ensure_dirs()
    
    # Load all cached races
    races = load_all_cached_races(date)
    print(f"Loaded {len(races)} races")
    
    total_horses = 0
    total_races = 0
    
    for race_data in races:
        track = race_data.get('track', '')
        race_num = race_data.get('raceNumber', 0)
        
        # Extract horses
        horses = []
        for r in race_data.get('runners', []):
            if r.get('scratched'):
                continue
            
            stats = r.get('stats', {})
            overall = stats.get('overall', {})
            track_stats = stats.get('track', {})
            distance = stats.get('distance', {})
            track_dist = stats.get('trackDistance', {})
            cond_stats = stats.get('conditions', {})
            first_up = stats.get('firstUp', {})
            second_up = stats.get('secondUp', {})
            
            # Get condition-specific stats
            cond_key = None
            for key in cond_stats:
                if cond_stats[key].get('starts', 0) > 0:
                    cond_key = key
                    break
            cond_data = cond_stats.get(cond_key, {}) if cond_key else {}
            
            horse = {
                'number': r['number'],
                'name': r.get('name', 'Unknown'),
                'jockey': r.get('jockey', ''),
                'trainer': r.get('trainer', ''),
                'weight': r.get('weight', 0),
                'barrier': r.get('barrier', 0),
                'age': r.get('age', 0),
                'sex': r.get('sex', ''),
                'sire': r.get('sire', ''),
                'dam': r.get('dam', ''),
                'form': r.get('form', ''),
                'last20_starts': r.get('last20Starts', ''),
                'career_prize_money': r.get('careerPrizeMoney', '$0'),
                'stats': {
                    'overall': {
                        'starts': overall.get('starts', 0),
                        'wins': overall.get('wins', 0),
                        'places': overall.get('places', 0),
                        'winPercent': overall.get('winPercent', 0),
                        'placePercent': overall.get('placePercent', 0),
                    },
                    'track': {
                        'starts': track_stats.get('starts', 0),
                        'wins': track_stats.get('wins', 0),
                        'winPercent': track_stats.get('winPercent', 0),
                        'placePercent': track_stats.get('placePercent', 0),
                    },
                    'distance': {
                        'starts': distance.get('starts', 0),
                        'wins': distance.get('wins', 0),
                        'winPercent': distance.get('winPercent', 0),
                        'placePercent': distance.get('placePercent', 0),
                    },
                    'trackDistance': {
                        'starts': track_dist.get('starts', 0),
                        'wins': track_dist.get('wins', 0),
                        'winPercent': track_dist.get('winPercent', 0),
                        'placePercent': track_dist.get('placePercent', 0),
                    },
                    'condition': {
                        'starts': cond_data.get('starts', 0),
                        'wins': cond_data.get('wins', 0),
                        'winPercent': cond_data.get('winPercent', 0),
                        'placePercent': cond_data.get('placePercent', 0),
                    },
                    'firstUp': {
                        'starts': first_up.get('starts', 0),
                        'wins': first_up.get('wins', 0),
                        'winPercent': first_up.get('winPercent', 0),
                    },
                    'secondUp': {
                        'starts': second_up.get('starts', 0),
                        'wins': second_up.get('wins', 0),
                        'winPercent': second_up.get('winPercent', 0),
                    },
                    'conditions': cond_stats,
                },
            }
            horses.append(horse)
        
        if not horses:
            continue
        
        # Generate race form guide
        race_guide = generate_race_form_guide(race_data, horses)
        race_file = FORM_GUIDES_DIR / f"{date}-{track}-R{race_num}.md"
        with open(race_file, 'w') as f:
            f.write(race_guide)
        
        total_races += 1
        total_horses += len(horses)
        
        # Generate individual horse form guides
        for horse in horses:
            horse_guide = generate_horse_form_guide(horse)
            horse_file = FORM_GUIDES_DIR / f"{date}-{horse['name'].replace(' ', '-')}.md"
            with open(horse_file, 'w') as f:
                f.write(horse_guide)
    
    print(f"Generated {total_races} race form guides")
    print(f"Generated {total_horses} horse form guides")
    print(f"Saved to: {FORM_GUIDES_DIR}")
    print("=" * 60)
    print("Complete!")


if __name__ == "__main__":
    main()