feat: UNSTRING解析增强 + 跨FD数值统一 + 文件I/O模块

cobol_testgen/output.py

@@ -1,8 +1,13 @@
 """输出层：JSON输出（按文件分组入出力 + 工作存储区分）"""
 
 import json
+import logging
 from pathlib import Path
 
+from . import file_io
+
+logger = logging.getLogger(__name__)
+
 
 _INVERSE_OP = {'>': '<=', '<': '>=', '=': '<>', '>=': '<', '<=': '>'}
 
@@ -120,7 +125,7 @@ def output_json(records, outpath, roles=None, fd_fields=None, field_to_fd=None,
 
 
 def output_input_files(records, outdir, stem, roles, fd_fields, field_to_fd, open_dir,
-                       term_types=None):
+                       term_types=None, data_fields=None, select_info=None):
     term_types = term_types or ['normal'] * len(records)
     input_fds = {}
     for fd_name, fds_set in fd_fields.items():
@@ -137,7 +142,8 @@ def output_input_files(records, outdir, stem, roles, fd_fields, field_to_fd, ope
 
     outdir.mkdir(parents=True, exist_ok=True)
 
-    for fd_name, fds_set in input_fds.items():
+    fd_items = list(input_fds.items())
+    for fd_idx, (fd_name, fds_set) in enumerate(fd_items):
         normals = []
         abends = []
         direction = (open_dir or {}).get(fd_name, '')
@@ -155,7 +161,53 @@ def output_input_files(records, outdir, stem, roles, fd_fields, field_to_fd, ope
                 else:
                     normals.append(fd_rec)
 
+        # 丢弃次要 FD 的最后 2 条记录，触发 EOF/不匹配路径
+        if fd_idx > 0 and normals:
+            normals = normals[:-2]
+
         if normals:
             _write_json(normals, outdir / f'{stem}_{fd_name}.json')
         if abends:
             _write_json(abends, outdir / f'{stem}_abend_{fd_name}.json')
+
+        if data_fields and select_info and normals:
+            assign_name = select_info.get(fd_name, {}).get('assign')
+            if assign_name:
+                bin_path = outdir / assign_name
+                name_to_field = {
+                    f['name']: f for f in data_fields
+                    if not f.get('is_88') and not f.get('is_filler')
+                    and f.get('pic')
+                }
+                field_dicts = []
+                seen = set()
+                for fname in fds_set:
+                    if fname in seen:
+                        continue
+                    seen.add(fname)
+                    fd = name_to_field.get(fname)
+                    if fd:
+                        field_dicts.append(fd)
+                if field_dicts:
+                    offsets = []
+                    offset = 0
+                    for fd in field_dicts:
+                        offsets.append(offset)
+                        offset += file_io.get_storage_length(fd)
+                    rec_len = offset
+                    if rec_len > 0:
+                        with open(bin_path, 'wb') as f:
+                            for rec in normals:
+                                buf = bytearray(rec_len)
+                                for fd, off in zip(field_dicts, offsets):
+                                    val = rec.get(fd['name'], '')
+                                    try:
+                                        packed = file_io.pack_value(val, fd)
+                                    except Exception as e:
+                                        logger.debug(f"pack_value failed for {fd['name']}: {e}")
+                                        slen = file_io.get_storage_length(fd)
+                                        packed = b'\x00' * slen
+                                    end = min(off + len(packed), rec_len)
+                                    buf[off:end] = packed[:end - off]
+                                f.write(bytes(buf))
+                        logger.info(f"  wrote {len(normals)} binary records to {bin_path}")