增加生成学校易耗品填写的转换程序

2026-08-31 18:25:35 +08:00
parent 44496a54e6
commit 4fbb4c36ed
19 changed files with 777 additions and 365 deletions
@@ -19,9 +19,7 @@ from .llm_extractor import (
    llm_query_text,
    load_cache,
    load_match_result,
-    merge_supplement_into_info,
    parse_json_response,
-    process_user_supplement,
 )

 __all__ = [
@@ -34,9 +32,7 @@ __all__ = [
    "load_cache",
    "load_match_result",
    "llm_query_text",
-    "merge_supplement_into_info",
    "parse_json_response",
-    "process_user_supplement",
    "EXTRACTION_PARALLEL_COUNT",
    "FILE_EVENTS_LOG",
    "SUPPORTED_EXTENSIONS",
@@ -115,6 +115,8 @@ def _build_summary(data: dict[str, Any]) -> dict[str, str]:
        "general": "普通发票",
        "payment": "支付记录",
        "application": "出差申请单",
+        "attachments": "其他附件",
+        "note": "易耗品单",
    }

    summary = {}
@@ -259,11 +261,12 @@ def extract_invoices(
        - 出差事前申请单 -> 单独存储，不参与匹配

    Returns:
-        (payment_records, applications, groups):
-            - payment_records: 支付记录列表（仅包含真实发票，不含申请单）
-            - applications: 出差事前申请单列表（单独存储，不参与支付匹配）
+        元组 (payment_records, applications, groups):
+            - payment_records: 支付记录列表（每项含关联发票 _matched_invoices）
+            - applications: 出差事前申请单列表
            - groups: 按文档类型分组的字典
-                {'travel': [差旅发票], 'general': [普通发票], 'application': [出差事前申请单]}
+                {'travel': [差旅发票], 'general': [普通发票], 'application': [出差事前申请单],
+                 'attachments': [其他附件]}

    Raises:
        ExtractionError: 当所有文件提取均失败时抛出，携带失败文件列表和原始错误。
@@ -280,13 +283,14 @@ def extract_invoices(
    all_files = _find_all_files(directory)
    if not all_files:
        log.warning("未找到支持的文件")
-        return [], [], {"travel": [], "general": [], "application": []}
+        return [], [], {"travel": [], "general": [], "application": [], "attachments": []}

    log.info(f"发现 {len(all_files)} 个文件")

    all_invoices = []
    all_cards = []
    applications = []
+    attachments = []
    # 记录失败文件及其错误信息
    failed_files: list[tuple[str, str]] = []

@@ -318,6 +322,9 @@ def extract_invoices(
            elif inv_type == "payment":
                all_cards.append(result)
                log.info(f"[{inv_type}] 已解析: {file_path.name}")
+            elif inv_type in ("attachments", "note"):
+                attachments.append(result)
+                log.info(f"[{inv_type}] 已解析: {file_path.name}")
            elif result.get("invoice_number"):
                all_invoices.append(result)
                log.info(f"[{inv_type}] 已解析: {file_path.name}")
@@ -334,7 +341,9 @@ def extract_invoices(
            details=error_details,
        )

-    log.info(f"分类结果: 发票 {len(all_invoices)} 张, 支付记录 {len(all_cards)} 条, 申请单 {len(applications)} 份")
+    log.info(
+        f"分类结果: 发票 {len(all_invoices)} 张, 支付记录 {len(all_cards)} 条, 申请单 {len(applications)} 份, 附件 {len(attachments)} 个"
+    )

    if not all_invoices:
        log.warning("未成功解析任何发票")
@@ -349,10 +358,12 @@ def extract_invoices(
    all_documents.extend(applications)

    groups = classify_invoice_batch(all_documents)
+    groups["attachments"] = attachments
    log.info(
        f"文档分类: 差旅发票 {len(groups['travel'])} 张, "
        f"普通发票 {len(groups['general'])} 张, "
-        f"出差事前申请单 {len(groups['application'])} 份"
+        f"出差事前申请单 {len(groups['application'])} 份, "
+        f"附件 {len(groups['attachments'])} 个"
    )

    return payment_records, applications, groups
@@ -43,7 +43,6 @@ from ...infra.documents.invoice import CACHE_DIR_NAME
 from ...infra.llm.prompt import (
    build_invoice_system_prompt,
    build_normal_info_system_prompt,
-    build_supplement_system_prompt,
    build_travel_info_system_prompt,
 )

@@ -386,6 +385,7 @@ def build_extraction_user_message(
    cache_map: dict[str, Any],
    match_result: dict[str, list[dict[str, Any]]],
    previous_analysis: dict[str, Any] | None = None,
+    user_supplement: str | None = None,
 ) -> str:
    """构建提取请求的用户消息（供 Agent 调度使用）。

@@ -393,6 +393,7 @@ def build_extraction_user_message(
        cache_map: 缓存数据映射。
        match_result: 匹配结果。
        previous_analysis: 上一轮 LLM 分析结果（可选，补充文件时传入作为历史上下文）。
+        user_supplement: 用户文字补充（可选，传入时作为修正上下文）。

    Returns:
        拼接好的用户消息字符串。
@@ -411,6 +412,9 @@ def build_extraction_user_message(
            + json.dumps(previous_analysis, ensure_ascii=False, indent=2)
        )

+    if user_supplement:
+        parts.append("【用户文字补充】用户补充了以下文字信息，请据此更新你的分析结果。\n" + user_supplement)
+
    if match_result:
        parts.append(
            "【发票与支付记录匹配结果】"
@@ -500,92 +504,3 @@ def extract_normal_info(
        包含报销说明、发票总数、总金额、支付方式、附件清单等字段的字典。
    """
    return _extract_info(source_dir, build_normal_info_system_prompt(), "普通发票")
-
-
-# ------------------------------------------------------------------
-# 用户补充信息处理
-# ------------------------------------------------------------------
-
-
-def process_user_supplement(
-    user_text: str,
-    extracted_info: dict[str, Any],
-    invoice_type: str,
-    source_dir: Path | None = None,
-) -> dict[str, Any]:
-    """让用户补充的文字信息通过 LLM 分析，返回需要更新的字段。
-
-    Args:
-        user_text: 用户输入的文字。
-        extracted_info: 当前已提取的报销信息。
-        invoice_type: "travel" 或 "normal"。
-        source_dir: 会话目录（可选，传入时启用 SSE 流式事件写入）。
-
-    Returns:
-        包含 updated_fields, changes, confidence, unparsed_info 的字典。
-    """
-    system_prompt = build_supplement_system_prompt()
-
-    parts = [
-        f"发票类型: {'差旅报销' if invoice_type == 'travel' else '普通报销'}",
-        "",
-        "以下是当前已提取的报销信息：",
-        json.dumps(extracted_info, ensure_ascii=False, indent=2),
-        "",
-        f"用户补充信息：{user_text}",
-        "",
-        "=== 请分析用户输入并返回需要更新的字段 ===",
-    ]
-    user_message = "\n".join(parts)
-
-    try:
-        response = llm_query_text(
-            system_prompt=system_prompt,
-            text=user_message,
-            reasoning_effort="low",
-            source_dir=source_dir,
-        )
-        result = parse_json_response(response)
-        log.info("LLM 补充信息分析完成")
-
-        return result
-    except Exception as e:
-        log.error("LLM 补充信息分析失败: %s", e)
-        return {
-            "updated_fields": {},
-            "changes": [],
-            "confidence": 0.0,
-            "unparsed_info": f"分析失败: {e}",
-        }
-
-
-def merge_supplement_into_info(
-    extracted_info: dict[str, Any],
-    updated_fields: dict[str, Any],
-) -> dict[str, Any]:
-    """将 LLM 返回的更新字段合并到已提取的信息中。
-
-    支持点号路径（如 basic_info.travel_purpose）表示嵌套更新。
-
-    Args:
-        extracted_info: 当前已提取的报销信息。
-        updated_fields: LLM 返回的需要更新的字段。
-
-    Returns:
-        更新后的报销信息。
-    """
-    import copy
-
-    result = copy.deepcopy(extracted_info)
-
-    for field_path, value in updated_fields.items():
-        parts = field_path.split(".")
-        current = result
-        for part in parts[:-1]:
-            if part not in current:
-                current[part] = {}
-            current = current[part]
-        current[parts[-1]] = value
-        log.info("更新字段 %s = %s", field_path, value)
-
-    return result
@@ -25,7 +25,7 @@ log = get_logger("validator")
 # 日期格式
 # ------------------------------------------------------------------

-DATE_PATTERN = re.compile(r"^\d{4}-\d{2}-\d{2}$")
+DATE_PATTERN = re.compile(r"^\d{4}-(?:0?[1-9]|1[0-2])-(?:0?[1-9]|[12]\d|3[01])$")


 def _is_valid_date(value: str) -> bool:
@@ -95,7 +95,7 @@ class ValidationConfig(TypedDict):
 # 配置加载
 # ------------------------------------------------------------------

-_CONFIG_PATH = Path(__file__).parent.parent.parent / "config" / "validation_rules.json"
+_CONFIG_PATH = Path(__file__).parent.parent.parent.parent / "config" / "validation_rules.json"
 _cached_rules: ValidationConfig | None = None