LLM-Driven Extraction of Unstructured Data — Built for API Deployments & ETL Pipeline Workflows
Python ai-agentsdata-engineeringdocument-aigenerative-aiidpjson-extractionllmmcp-serverocrpdf-extractionprompt-engineeringstructured-output
from collections.abc import Iterable
from account_v2.models import User
def batch_resolve_user_ids(user_ids: Iterable[str]) -> dict[str, dict[str, str]]:
"""Batch resolve user IDs to display info in a single query.
Accepts stringified integer PKs (as stored in CharField fields like
HITLQueue.reviewer_id / approver_id) and returns a lookup dict
mapping each string PK to display info.
Args:
user_ids: Iterable of stringified user PKs to resolve.
Returns:
dict mapping each string PK to {'name': str, 'email': str}.
PKs that cannot be resolved (non-existent users or invalid IDs)
are omitted from the result; callers should use `.get()` for safe access.
"""
pk_set = set()
for uid in user_ids or []:
try:
pk_set.add(int(uid))
except (ValueError, TypeError):
continue
if not pk_set:
return {}
lookup = {}
for user in User.objects.filter(pk__in=pk_set):
display_name = (
user.get_full_name()
or (user.email.split("@")[0] if user.email else None)
or user.username
)
info = {"name": display_name, "email": user.email or ""}
lookup[str(user.pk)] = info
return lookup