98fa16bfc3
Full-stack Amazon AI Transcreation Platform with: - FastAPI backend (async, PostgreSQL, Redis, Celery) with 11 DB tables - JWT auth (SSO-ready abstract provider pattern) - 6-agent pipeline orchestrator with deterministic modules - Next.js 14 frontend with Amazon branding (Ember fonts, orange/dark theme) - Job wizard, monitoring HUD, output review, admin screens - 154 TM/reference files imported, 12 locales configured - Docker Compose for all services Agents 2-5 (TM retrieval, ranker, transcreator, compliance) are stubs pending Phase 3 LLM integration. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
112 lines
4 KiB
Python
112 lines
4 KiB
Python
"""Agent 6: Formatter
|
|
|
|
Generates the output xlsx file and builds output row data.
|
|
This agent is deterministic (no LLM call).
|
|
"""
|
|
|
|
from datetime import datetime, timezone
|
|
from typing import Any
|
|
from uuid import uuid4
|
|
|
|
from app.pipeline.agents.base import BaseAgent
|
|
from app.pipeline.contracts import PipelineContext
|
|
from app.pipeline.modules.excel_writer import generate_output_xlsx
|
|
from app.config import settings
|
|
|
|
|
|
class Agent6Formatter(BaseAgent):
|
|
"""Formats pipeline output into xlsx and structured data."""
|
|
|
|
name = "agent_6_formatter"
|
|
description = "Generates output xlsx and structured output rows"
|
|
|
|
def __init__(self, output_dir: str | None = None) -> None:
|
|
self.output_dir = output_dir or settings.STORAGE_ROOT
|
|
|
|
def get_system_prompt(self) -> str:
|
|
return "" # No LLM call
|
|
|
|
def build_user_message(self, context: PipelineContext) -> str:
|
|
return "" # No LLM call
|
|
|
|
def parse_response(self, response: str, context: PipelineContext) -> Any:
|
|
return None # No LLM call
|
|
|
|
async def run(self, context: PipelineContext) -> PipelineContext:
|
|
"""Generate output xlsx and return updated context with file path.
|
|
|
|
Returns the context as-is (output rows are built by the orchestrator
|
|
for database persistence). The xlsx file is written to disk.
|
|
"""
|
|
job_id = context.job_params.job_id
|
|
locale_code = context.job_params.locale_code
|
|
|
|
# Build source lines for excel
|
|
source_lines_data = [
|
|
{
|
|
"row_order": sl.row_order,
|
|
"en_gb": sl.en_gb,
|
|
"copy_type": sl.copy_type or "",
|
|
}
|
|
for sl in context.source_lines
|
|
]
|
|
|
|
# Build output rows for excel
|
|
output_rows_data = []
|
|
for i, draft in enumerate(context.draft_outputs):
|
|
row: dict[str, Any] = {
|
|
"row_order": i + 1,
|
|
"line_id": draft.line_id,
|
|
"option_1": draft.option_1.text if draft.option_1 else "",
|
|
"backtranslation_1": draft.option_1.backtranslation if draft.option_1 else "",
|
|
"rationale_1": draft.option_1.rationale if draft.option_1 else "",
|
|
}
|
|
if draft.option_2:
|
|
row["option_2"] = draft.option_2.text
|
|
row["backtranslation_2"] = draft.option_2.backtranslation
|
|
row["rationale_2"] = draft.option_2.rationale
|
|
if draft.option_3:
|
|
row["option_3"] = draft.option_3.text
|
|
row["backtranslation_3"] = draft.option_3.backtranslation
|
|
row["rationale_3"] = draft.option_3.rationale
|
|
|
|
output_rows_data.append(row)
|
|
|
|
# Build summary
|
|
compliance_counts = {"high": 0, "moderate": 0, "low": 0}
|
|
for ranking in context.ranking_declarations:
|
|
tier = ranking.confidence_tier
|
|
if tier in compliance_counts:
|
|
compliance_counts[tier] += 1
|
|
|
|
summary = {
|
|
"job_id": job_id,
|
|
"campaign_name": context.job_params.campaign_name,
|
|
"locale_code": locale_code,
|
|
"channel": context.job_params.channel,
|
|
"programme": context.job_params.programme,
|
|
"total_source_lines": len(context.source_lines),
|
|
"total_output_rows": len(output_rows_data),
|
|
"high_confidence": compliance_counts["high"],
|
|
"moderate_confidence": compliance_counts["moderate"],
|
|
"low_confidence": compliance_counts["low"],
|
|
"total_tokens": 0,
|
|
"estimated_cost": 0.0,
|
|
"agent_version": "1.0.0",
|
|
"generated_at": datetime.now(timezone.utc).isoformat(),
|
|
}
|
|
|
|
# Generate xlsx
|
|
output_path = (
|
|
f"{self.output_dir}/jobs/{job_id}/output/"
|
|
f"{locale_code}_{job_id}_output.xlsx"
|
|
)
|
|
|
|
generate_output_xlsx(
|
|
output_path=output_path,
|
|
source_lines=source_lines_data,
|
|
output_rows=output_rows_data,
|
|
summary=summary,
|
|
)
|
|
|
|
return context
|