[birds] Add birding csv importer

vrobbler/apps/birds/importer.py

@@ -0,0 +1,152 @@
+import csv
+import logging
+import re
+from collections import defaultdict
+from datetime import datetime
+from zoneinfo import ZoneInfo
+
+from django.contrib.auth import get_user_model
+
+from birds.models import Bird, BirdSightingEntry, BirdSightingLogData, BirdingLocation
+from scrobbles.models import Scrobble
+
+logger = logging.getLogger(__name__)
+User = get_user_model()
+
+LOCATION_COORDS_RE = re.compile(r"\(([\d\.\-]+),\s*([\d\.\-]+)\)")
+DURATION_RE = re.compile(r"(\d+)\s*minute")
+
+
+def parse_duration(duration_str):
+    if not duration_str:
+        return None
+    match = DURATION_RE.search(duration_str)
+    if match:
+        return int(match.group(1))
+    return None
+
+
+def parse_coords(location_str):
+    match = LOCATION_COORDS_RE.search(location_str)
+    if match:
+        return float(match.group(1)), float(match.group(2))
+    return None, None
+
+
+def parse_timestamp(date_str, time_str):
+    try:
+        dt = datetime.strptime(f"{date_str} {time_str}", "%B %d, %Y %I:%M %p")
+        return dt.replace(tzinfo=ZoneInfo("UTC"))
+    except (ValueError, TypeError):
+        try:
+            dt = datetime.strptime(date_str, "%B %d, %Y")
+            return dt.replace(tzinfo=ZoneInfo("UTC"))
+        except (ValueError, TypeError):
+            logger.warning(f"Could not parse date/time: {date_str} {time_str}")
+            return None
+
+
+def import_birding_csv(file_path, user_id):
+    user = User.objects.get(id=user_id)
+    new_scrobbles = []
+
+    with open(file_path, newline="", encoding="utf-8-sig") as f:
+        reader = csv.DictReader(f)
+        rows = list(reader)
+
+    groups = defaultdict(list)
+    for row in rows:
+        key = (
+            row.get("Location", "").strip(),
+            row.get("Observation Date", "").strip(),
+            row.get("Start Time", "").strip(),
+        )
+        groups[key].append(row)
+
+    for (location_str, date_str, time_str), sighting_rows in groups.items():
+        if not location_str:
+            logger.warning("Skipping rows with no location")
+            continue
+
+        timestamp = parse_timestamp(date_str, time_str)
+        if not timestamp:
+            continue
+
+        location_title = LOCATION_COORDS_RE.sub("", location_str).strip().rstrip(",").strip()
+        if not location_title:
+            location_title = location_str
+
+        location = BirdingLocation.find_or_create(location_title)
+        lat, lon = parse_coords(location_str)
+        if lat and lon and not location.geo_location:
+            from locations.models import GeoLocation
+
+            geo, _ = GeoLocation.objects.get_or_create(
+                lat=round(lat, 6),
+                lon=round(lon, 6),
+                defaults={"altitude": None},
+            )
+            location.geo_location = geo
+            location.save(update_fields=["geo_location"])
+
+        first_row = sighting_rows[0]
+        duration_minutes = parse_duration(first_row.get("Duration", ""))
+        party_size = first_row.get("Party Size", "").strip()
+
+        birds_data = []
+        for row in sighting_rows:
+            species = row.get("Species", "").strip()
+            if not species:
+                continue
+            count_str = row.get("Count", "1").strip()
+            try:
+                count = int(count_str) if count_str else 1
+            except ValueError:
+                count = 1
+            details = row.get("Details", "").strip()
+
+            bird = Bird.find_or_create(species)
+            entry = BirdSightingEntry(
+                bird_id=bird.id, quantity=count, sighting_notes=details or None
+            )
+            birds_data.append(entry.asdict)
+
+        logdata = BirdSightingLogData(
+            birds=birds_data,
+            duration_minutes=duration_minutes,
+        )
+
+        notes = []
+        if party_size:
+            notes.append(f"Party size: {party_size}")
+        checklist_flag = first_row.get("Complete Checklist", "").strip()
+        if checklist_flag:
+            notes.append(f"Complete checklist: {checklist_flag}")
+
+        log_dict = logdata.asdict
+        if notes:
+            log_dict["notes"] = notes
+
+        scrobble = Scrobble(
+            user=user,
+            timestamp=timestamp,
+            source="Birding CSV Import",
+            birding_location=location,
+            log=log_dict,
+            played_to_completion=True,
+            in_progress=False,
+            media_type=Scrobble.MediaType.BIRDING_LOCATION,
+        )
+        existing = Scrobble.objects.filter(
+            timestamp=timestamp,
+            birding_location=location,
+            user=user,
+        ).first()
+        if existing:
+            logger.debug(f"Skipping existing scrobble for {location}")
+            continue
+        new_scrobbles.append(scrobble)
+
+    created = Scrobble.objects.bulk_create(new_scrobbles)
+    logger.info(f"Created {len(created)} birding scrobbles")
+    return created