fix docker scraping

improve batch sync

add postgres upsert

fix sync locking

improve listing sync

speed up scraper

clean up project

prepare for github

update docker setup
This commit is contained in:
qananasikq
2026-04-13 16:35:08 +03:00
parent b9557922ed
commit a4c93a1df1
30 changed files with 2239 additions and 624 deletions

View File

@@ -1,7 +1,7 @@
import json
import re
from pathlib import Path
from typing import Any, Iterable
from typing import Any, Callable, Iterable
def save_to_json(data: Any, filename: str | Path) -> None:
@@ -17,7 +17,7 @@ def first_non_empty(values: Iterable[Any]) -> Any | None:
return None
# Regex для VIN, lot, price.
# Регулярные выражения для VIN, lot и price.
VIN_RE = re.compile(r"\b([A-HJ-NPR-Z0-9]{17})\b", re.IGNORECASE)
LOT_RE = re.compile(r"\b(\d{7,10})\b")
PRICE_RE = re.compile(r"\$\s?([\d,]+(?:\.\d{1,2})?)")
@@ -37,3 +37,36 @@ def deep_find_key(obj, target_keys: set[str], max_depth: int = 64, _depth: int =
for item in obj:
found.extend(deep_find_key(item, target_keys, max_depth=max_depth, _depth=_depth + 1))
return found
def deep_find_all_keys(
payloads: list,
field_map: dict[str, set[str]],
max_depth: int = 64,
) -> dict[str, list]:
"""Извлекает все нужные поля за один проход по JSON."""
# Готовим обратную карту: нормализованный ключ -> имя поля.
reverse: dict[str, str] = {}
for field_name, keys in field_map.items():
for k in keys:
reverse[k.lower()] = field_name
result: dict[str, list] = {f: [] for f in field_map}
def _recurse(obj: Any, depth: int) -> None:
if depth >= max_depth:
return
if isinstance(obj, dict):
for k, v in obj.items():
field = reverse.get(k.lower())
if field is not None:
result[field].append(v)
_recurse(v, depth + 1)
elif isinstance(obj, list):
for item in obj:
_recurse(item, depth + 1)
for payload in payloads:
_recurse(payload, 0)
return result