Remove flaresolverr and optimize refresh upserts
This commit is contained in:
@@ -52,12 +52,6 @@
|
|||||||
MOBILEDE_BOOTSTRAP_FULL_SCAN_ENABLED: ${MOBILEDE_BOOTSTRAP_FULL_SCAN_ENABLED:-true}
|
MOBILEDE_BOOTSTRAP_FULL_SCAN_ENABLED: ${MOBILEDE_BOOTSTRAP_FULL_SCAN_ENABLED:-true}
|
||||||
MOBILEDE_INCREMENTAL_AFTER_BOOTSTRAP: ${MOBILEDE_INCREMENTAL_AFTER_BOOTSTRAP:-true}
|
MOBILEDE_INCREMENTAL_AFTER_BOOTSTRAP: ${MOBILEDE_INCREMENTAL_AFTER_BOOTSTRAP:-true}
|
||||||
MOBILEDE_INCREMENTAL_PAGE_WINDOW: ${MOBILEDE_INCREMENTAL_PAGE_WINDOW:-1}
|
MOBILEDE_INCREMENTAL_PAGE_WINDOW: ${MOBILEDE_INCREMENTAL_PAGE_WINDOW:-1}
|
||||||
MOBILEDE_FLARESOLVERR_ENABLED: ${MOBILEDE_FLARESOLVERR_ENABLED:-false}
|
|
||||||
MOBILEDE_FLARESOLVERR_URL: ${MOBILEDE_FLARESOLVERR_URL:-http://flaresolverr:8191/v1}
|
|
||||||
MOBILEDE_FLARESOLVERR_TIMEOUT_SECONDS: ${MOBILEDE_FLARESOLVERR_TIMEOUT_SECONDS:-120}
|
|
||||||
MOBILEDE_FLARESOLVERR_MAX_TIMEOUT_MS: ${MOBILEDE_FLARESOLVERR_MAX_TIMEOUT_MS:-60000}
|
|
||||||
MOBILEDE_FLARESOLVERR_SESSION: ${MOBILEDE_FLARESOLVERR_SESSION:-}
|
|
||||||
MOBILEDE_FLARESOLVERR_STATUSES: ${MOBILEDE_FLARESOLVERR_STATUSES:-403,429,503}
|
|
||||||
IAAI_STARTUP_SYNC_ENABLED: ${IAAI_STARTUP_SYNC_ENABLED:-true}
|
IAAI_STARTUP_SYNC_ENABLED: ${IAAI_STARTUP_SYNC_ENABLED:-true}
|
||||||
IAAI_INTER_BATCH_DELAY_SECONDS: ${IAAI_INTER_BATCH_DELAY_SECONDS:-0}
|
IAAI_INTER_BATCH_DELAY_SECONDS: ${IAAI_INTER_BATCH_DELAY_SECONDS:-0}
|
||||||
IAAI_FAIL_RATE_THRESHOLD: ${IAAI_FAIL_RATE_THRESHOLD:-0.9}
|
IAAI_FAIL_RATE_THRESHOLD: ${IAAI_FAIL_RATE_THRESHOLD:-0.9}
|
||||||
@@ -150,23 +144,6 @@ services:
|
|||||||
max-size: "10m"
|
max-size: "10m"
|
||||||
max-file: "5"
|
max-file: "5"
|
||||||
|
|
||||||
flaresolverr:
|
|
||||||
image: ghcr.io/flaresolverr/flaresolverr:latest
|
|
||||||
container_name: mobilede-flaresolverr
|
|
||||||
restart: unless-stopped
|
|
||||||
environment:
|
|
||||||
LOG_LEVEL: ${FLARESOLVERR_LOG_LEVEL:-info}
|
|
||||||
LOG_HTML: ${FLARESOLVERR_LOG_HTML:-false}
|
|
||||||
CAPTCHA_SOLVER: ${FLARESOLVERR_CAPTCHA_SOLVER:-none}
|
|
||||||
TZ: ${TZ:-UTC}
|
|
||||||
ports:
|
|
||||||
- "127.0.0.1:${MOBILEDE_HOST_FLARESOLVERR_PORT:-8191}:8191"
|
|
||||||
logging:
|
|
||||||
driver: json-file
|
|
||||||
options:
|
|
||||||
max-size: "10m"
|
|
||||||
max-file: "5"
|
|
||||||
|
|
||||||
# Миграции.
|
# Миграции.
|
||||||
migrate:
|
migrate:
|
||||||
<<: *app-service
|
<<: *app-service
|
||||||
|
|||||||
@@ -248,6 +248,8 @@ class MobileDeMapper:
|
|||||||
return "US"
|
return "US"
|
||||||
if text in {"CA", "CANADA"}:
|
if text in {"CA", "CANADA"}:
|
||||||
return "CA"
|
return "CA"
|
||||||
|
if text in {"IT", "ITALY", "ITALIA"}:
|
||||||
|
return "IT"
|
||||||
if text in {"JP", "JAPAN"}:
|
if text in {"JP", "JAPAN"}:
|
||||||
return "JP"
|
return "JP"
|
||||||
if text in {"KR", "KOREA", "SOUTH KOREA"}:
|
if text in {"KR", "KOREA", "SOUTH KOREA"}:
|
||||||
|
|||||||
@@ -1,4 +1,5 @@
|
|||||||
import logging
|
import logging
|
||||||
|
import os
|
||||||
from contextlib import contextmanager
|
from contextlib import contextmanager
|
||||||
from datetime import datetime, timezone
|
from datetime import datetime, timezone
|
||||||
from typing import Any, Iterator
|
from typing import Any, Iterator
|
||||||
@@ -12,6 +13,7 @@ from .models import Base, Car, Image, SyncRun
|
|||||||
from .schemas import CarRecord
|
from .schemas import CarRecord
|
||||||
|
|
||||||
logger = logging.getLogger("MOBILEDE_scraper.db")
|
logger = logging.getLogger("MOBILEDE_scraper.db")
|
||||||
|
MOBILEDE_SKIP_IMAGES_FOR_UPDATED = os.getenv("MOBILEDE_SKIP_IMAGES_FOR_UPDATED", "1").strip().lower() in {"1", "true", "yes", "on"}
|
||||||
|
|
||||||
|
|
||||||
CAR_DB_FIELDS = {
|
CAR_DB_FIELDS = {
|
||||||
@@ -264,17 +266,19 @@ class PersistenceService:
|
|||||||
images = [image.model_dump(mode="python") for image in record.images]
|
images = [image.model_dump(mode="python") for image in record.images]
|
||||||
car_by_id = existing_by_id.get(record.origin_id)
|
car_by_id = existing_by_id.get(record.origin_id)
|
||||||
car_by_url = existing_by_url.get(record.origin_url)
|
car_by_url = existing_by_url.get(record.origin_url)
|
||||||
entry: dict[str, object] = {"record": record, "images": images, "car_id": None}
|
entry: dict[str, object] = {"record": record, "images": images, "car_id": None, "action": "inserted"}
|
||||||
|
|
||||||
if car_by_url is not None and car_by_url.origin_id != record.origin_id and car_by_id is None:
|
if car_by_url is not None and car_by_url.origin_id != record.origin_id and car_by_id is None:
|
||||||
for key, value in payload.items():
|
for key, value in payload.items():
|
||||||
setattr(car_by_url, key, value)
|
setattr(car_by_url, key, value)
|
||||||
car_by_url.last_seen_at = record.last_seen_at
|
car_by_url.last_seen_at = record.last_seen_at
|
||||||
entry["car_id"] = int(car_by_url.id)
|
entry["car_id"] = int(car_by_url.id)
|
||||||
|
entry["action"] = "updated"
|
||||||
updated += 1
|
updated += 1
|
||||||
else:
|
else:
|
||||||
upsert_payloads.append(payload)
|
upsert_payloads.append(payload)
|
||||||
if car_by_id is not None:
|
if car_by_id is not None:
|
||||||
|
entry["action"] = "updated"
|
||||||
updated += 1
|
updated += 1
|
||||||
else:
|
else:
|
||||||
inserted += 1
|
inserted += 1
|
||||||
@@ -306,7 +310,10 @@ class PersistenceService:
|
|||||||
|
|
||||||
for entry in entries:
|
for entry in entries:
|
||||||
car_id = int(entry["car_id"])
|
car_id = int(entry["car_id"])
|
||||||
|
action = str(entry.get("action") or "updated")
|
||||||
images = entry["images"]
|
images = entry["images"]
|
||||||
|
if MOBILEDE_SKIP_IMAGES_FOR_UPDATED and action == "updated":
|
||||||
|
continue
|
||||||
new_image_urls = {
|
new_image_urls = {
|
||||||
str(img.get("fullres_image", ""))
|
str(img.get("fullres_image", ""))
|
||||||
for img in images
|
for img in images
|
||||||
|
|||||||
@@ -16,7 +16,7 @@ BODY_TYPE_ENUM_VALUES = (
|
|||||||
"OTHER",
|
"OTHER",
|
||||||
"NA",
|
"NA",
|
||||||
)
|
)
|
||||||
COUNTRY_ENUM_VALUES = ("JP", "KR", "US", "CA", "DE", "NA")
|
COUNTRY_ENUM_VALUES = ("JP", "KR", "US", "CA", "DE", "IT", "NA")
|
||||||
ORIGIN_ENUM_VALUES = (
|
ORIGIN_ENUM_VALUES = (
|
||||||
"MOBILEDE",
|
"MOBILEDE",
|
||||||
"MOBILE_DE",
|
"MOBILE_DE",
|
||||||
|
|||||||
Reference in New Issue
Block a user