69a673e6c6
Chat 访谈 - 新增 persona 系统（default / warm_listener / curious_guide）与 background_voice 语气层 - 回复长度由 compute_reply_plan 统一决策（brief / standard / expanded），融合信息密度启发式 - 输入净稿（input_normalize）：编排层可选 rules/llm 归一用户口语后再喂模型与记忆检索 - 记忆证据注入：按用户话检索 memory evidence 并注入 prompt Memoir 回忆录 - 口述归一（oral_normalize）：segment 原文保留，story 管线取派生净稿作叙事输入 - segment 入队批次门闸：累计字数 + 最长等待秒数，减少零碎提交 - fidelity_check / prompts / narrative_agent 微调 - Alembic 0005：清理跨章节 story 外键 Infra - Dockerfile 加入 ffmpeg - pyproject.toml 新增依赖并同步 uv.lock - .env.example / .env.production 补全新配置项 Tests - 新增 test_background_voice、test_chat_input_normalize、test_experience_regressions - 扩展 test_interview_prompts、test_interview_reply_length、test_story_route_oral_invariant Made-with: Cursor
108 lines
3.2 KiB
Python
108 lines
3.2 KiB
Python
"""JSON 载荷解析、证据格式化、Story 批量规划校验（纯函数）。"""
|
|
|
|
import pytest
|
|
|
|
from app.agents.chat.reply_limits import truncate_chat_segments
|
|
|
|
from app.agents.memoir.classification_agent import _normalize_llm_category
|
|
from app.agents.memoir.prompts import format_evidence_chunks_for_prompt
|
|
from app.features.memory.evidence_format import (
|
|
format_evidence_chunks_for_prompt as format_evidence_from_memory,
|
|
)
|
|
from app.agents.memoir.story_route_agent import (
|
|
StoryBatchPlan,
|
|
StoryBatchPlanUnit,
|
|
validate_story_batch_plan,
|
|
)
|
|
from app.features.memoir.memoir_images.json_payload import extract_json_payload
|
|
|
|
|
|
def test_extract_json_payload_strips_markdown_fence() -> None:
|
|
raw = """```json
|
|
{"a": 1}
|
|
```"""
|
|
assert '"a"' in extract_json_payload(raw)
|
|
|
|
|
|
def test_normalize_llm_category_strips_quotes() -> None:
|
|
assert _normalize_llm_category('"childhood"') == "childhood"
|
|
assert _normalize_llm_category("`beliefs`") == "beliefs"
|
|
|
|
|
|
def test_format_evidence_chunks_includes_timeline() -> None:
|
|
ev = {
|
|
"relevant_chunks": [{"content": "chunk1"}],
|
|
"relevant_facts": [
|
|
{"subject": "我", "predicate": "生于", "object_json": "1950"}
|
|
],
|
|
"timeline_hints": [
|
|
{
|
|
"id": "1",
|
|
"event_year": 1977,
|
|
"event_date": None,
|
|
"title": "恢复高考",
|
|
"description": "参加了考试",
|
|
}
|
|
],
|
|
"relevant_summaries": [],
|
|
"relevant_stories": [],
|
|
}
|
|
out = format_evidence_chunks_for_prompt(ev)
|
|
assert "chunk1" in out
|
|
assert "1950" in out or "生于" in out
|
|
assert "1977" in out or "恢复高考" in out
|
|
assert format_evidence_from_memory(ev) == out
|
|
|
|
|
|
def test_validate_story_batch_plan_ok() -> None:
|
|
ordered = ["s1", "s2"]
|
|
plan = StoryBatchPlan(
|
|
units=[
|
|
StoryBatchPlanUnit(
|
|
segment_ids=["s1", "s2"],
|
|
decision="new_story",
|
|
target_story_id=None,
|
|
new_story_title="标题",
|
|
reason=None,
|
|
)
|
|
]
|
|
)
|
|
ok, err = validate_story_batch_plan(ordered, plan, valid_story_ids=set())
|
|
assert ok is True
|
|
assert err is None
|
|
|
|
|
|
def test_truncate_chat_segments() -> None:
|
|
out = truncate_chat_segments(
|
|
["a" * 300, "b"],
|
|
max_segments=2,
|
|
max_chars_per_segment=220,
|
|
)
|
|
assert out[0] == "a" * 219 + "…"
|
|
assert len(out[0]) == 220
|
|
assert out[1] == "b"
|
|
|
|
|
|
def test_validate_story_batch_plan_duplicate_segment() -> None:
|
|
plan = StoryBatchPlan(
|
|
units=[
|
|
StoryBatchPlanUnit(
|
|
segment_ids=["s1"],
|
|
decision="new_story",
|
|
target_story_id=None,
|
|
new_story_title="A",
|
|
reason=None,
|
|
),
|
|
StoryBatchPlanUnit(
|
|
segment_ids=["s1"],
|
|
decision="new_story",
|
|
target_story_id=None,
|
|
new_story_title="B",
|
|
reason=None,
|
|
),
|
|
]
|
|
)
|
|
ok, err = validate_story_batch_plan(["s1", "s1"], plan, valid_story_ids=set())
|
|
assert ok is False
|
|
assert err == "duplicate_segment"
|