init: cobol-java migration verification platform v3 (42 tests, JCL module)

DESIGN.md

@@ -0,0 +1,40 @@
+# DESIGN.md — verify-cli Web UI
+
+## Aesthetic
+Terminal Developer Tool — 命令行工具的可视化包装。暗底、等宽、最小装饰。
+用户的第一反应应该是"这是我的终端，只不过多了个表单"。
+
+## Typography
+- Primary: SF Mono / Fira Code / Cascadia Code / Consolas (等宽字体堆栈)
+- 全部使用系统原生字体，零外部依赖
+- 层级差异通过字号和颜色区分，不换字体家族
+
+## Color
+| Token | Hex | Usage |
+|-------|-----|-------|
+| bg | #0a0e14 | 页面底色 |
+| panel | #12171f | 卡片/表单容器 |
+| border | #1f2937 | 分割线 |
+| text | #b2becd | 正文 |
+| dim | #5c6e80 | 标签/次要信息 |
+| accent | #39bae6 | 链接/操作色 |
+| green | #7fd962 | 成功状态 |
+| red | #f26d78 | 错误状态 |
+| yellow | #ffad66 | 等待中状态 |
+
+## Layout
+- 单栏，最大宽度 680px
+- Header → Sections → Footer 的垂直流
+- 表单使用 CSS Grid 双列布局
+- 结果页 Section 分离 Summary 和 Field Results
+
+## Spacing
+- 页面 padding: 3rem 1.5rem
+- Section 内 padding: 1.5rem, 间距 1rem
+- 表单 label 间距: .75rem
+
+## Decisions
+- 不使用 Jinja2（3.1+ 与 Starlette 不兼容），改用字符串替换
+- 不使用任何 CSS 框架，纯手写 CSS variables
+- 不使用 emoji 或装饰图标，状态通过颜色边框表达
+- 无紫色渐变、无 3 列 icon grid、无居中布局、无装饰性波浪 — AI slop 反模式检查通过

config.py

@@ -1,48 +0,0 @@
-from dataclasses import dataclass, field
-from pathlib import Path
-
-
-@dataclass
-class Config:
-    project_name: str = ""
-    copybook_paths: list = field(default_factory=lambda: ["./copybooks"])
-    dialect: str = "ibm"
-    llm_model: str = "gpt-4o-mini"
-    llm_timeout: int = 15
-    llm_cache_dir: str = ".cache/llm"
-    coverage_default: str = "boundary"
-    rounding_mode: str = "TRUNCATE"
-    tolerance: float = 0.01
-    runner_mode: str = "native"
-    spark_master: str = "local[*]"
-    spark_input_format: str = "json"
-    num_records: int = 1000
-    branch_pass: float = 0.80
-    max_llm_cost: float = 0.50
-
-    @classmethod
-    def from_toml(cls, path="aurak.toml"):
-        import tomllib
-        try:
-            with open(path, "rb") as f:
-                d = tomllib.load(f)
-        except:
-            return cls()
-        c = cls()
-        p = d.get("project", {})
-        c.project_name = p.get("name", "")
-        c.copybook_paths = p.get("copybook_paths", c.copybook_paths)
-        c.dialect = p.get("dialect", "ibm")
-        ll = d.get("llm", {})
-        c.llm_model = ll.get("model", c.llm_model)
-        co = d.get("coverage", {})
-        c.coverage_default = co.get("default_target", "boundary")
-        cp = d.get("comparison", {})
-        c.rounding_mode = cp.get("rounding_mode", "TRUNCATE")
-        c.tolerance = cp.get("default_tolerance", c.tolerance)
-        r = d.get("runner", {})
-        c.runner_mode = r.get("mode", "native")
-        s = d.get("spark", {})
-        c.spark_master = s.get("master", "local[*]")
-        c.num_records = s.get("num_records", c.num_records)
-        return c

config/__init__.py

@@ -0,0 +1,49 @@
+from dataclasses import dataclass, field
+from pathlib import Path
+from .mapping import MappingConfig, FieldMapping
+
+
+@dataclass
+class Config:
+    project_name: str = ""
+    copybook_paths: list = field(default_factory=lambda: ["./copybooks"])
+    dialect: str = "ibm"
+    llm_model: str = "gpt-4o-mini"
+    llm_timeout: int = 15
+    llm_cache_dir: str = ".cache/llm"
+    coverage_default: str = "boundary"
+    rounding_mode: str = "TRUNCATE"
+    tolerance: float = 0.01
+    runner_mode: str = "native"
+    spark_master: str = "local[*]"
+    spark_input_format: str = "json"
+    num_records: int = 1000
+    branch_pass: float = 0.80
+    max_llm_cost: float = 0.50
+
+    @classmethod
+    def from_toml(cls, path="aurak.toml"):
+        import tomllib
+        try:
+            with open(path, "rb") as f:
+                d = tomllib.load(f)
+        except:
+            return cls()
+        c = cls()
+        p = d.get("project", {})
+        c.project_name = p.get("name", "")
+        c.copybook_paths = p.get("copybook_paths", c.copybook_paths)
+        c.dialect = p.get("dialect", "ibm")
+        ll = d.get("llm", {})
+        c.llm_model = ll.get("model", c.llm_model)
+        co = d.get("coverage", {})
+        c.coverage_default = co.get("default_target", "boundary")
+        cp = d.get("comparison", {})
+        c.rounding_mode = cp.get("rounding_mode", "TRUNCATE")
+        c.tolerance = cp.get("default_tolerance", c.tolerance)
+        r = d.get("runner", {})
+        c.runner_mode = r.get("mode", "native")
+        s = d.get("spark", {})
+        c.spark_master = s.get("master", "local[*]")
+        c.num_records = s.get("num_records", c.num_records)
+        return c

data/diff_result.py

@@ -30,6 +30,7 @@ class VerificationRun:
     branch_rate: float = 0.0
     llm_cost: float = 0.0
     report_path: str = ""
+    debug: dict = field(default_factory=dict)
 
     def __post_init__(self):
         if not self.timestamp:

jcl/executor.py

@@ -0,0 +1,101 @@
+"""JCL Executor — executes parsed JCL steps using platform components."""
+from pathlib import Path
+from jcl.parser import Job, JobStep, CondParam, COND_OPS
+
+
+class JclExecutor:
+    """Executes JCL Job steps. Uses platform CobolRunner for COBOL programs."""
+
+    def __init__(self, root_dir: str, cobol_dir: str, copybook_dir: str):
+        self.root_dir = Path(root_dir)
+        self.cobol_dir = Path(cobol_dir)
+        self.copybook_dir = Path(copybook_dir)
+        self.bin_dir = self.root_dir / "bin"
+        self.bin_dir.mkdir(exist_ok=True)
+        self.last_rc: int = 0
+        self.step_rcs: dict[str, int] = {}
+        self.results: dict[str, dict] = {}
+
+    def run(self, job: Job) -> int:
+        for step in job.steps:
+            rc = self._execute_step(step)
+            self.last_rc = rc
+            self.step_rcs[step.step_name] = rc
+
+        return self.last_rc
+
+    def _execute_step(self, step: JobStep) -> int:
+        # COND check
+        if step.cond and not self._check_cond(step.cond):
+            self.results[step.step_name] = {"status": "SKIPPED", "cond": str(step.cond)}
+            return 0
+
+        prog = step.program.upper()
+        if prog == "SORT":
+            return self._run_sort(step)
+
+        # Use platform's CobolRunner
+        import sys
+        sys.path.insert(0, str(Path(__file__).parent.parent))
+
+        from runners.cobol_runner import CobolRunner
+
+        cbl_path = self.cobol_dir / f"{step.program}.cbl"
+        runner = CobolRunner()
+        build = runner.compile(str(cbl_path))
+        if not build.success:
+            self.results[step.step_name] = {"status": "COMPILE_ERROR", "log": build.log[-200:]}
+            return 8
+
+        # Map DD entries to environment variables
+        env_in = {}
+        env_out = {}
+        for dd in step.dd_entries:
+            name = dd.dd_name.upper()
+            if dd.dsn:
+                path = self._resolve_path(dd.dsn)
+                if name in ("TRANSIN", "MEMBER", "VALIDIN", "RATE", "BILLING"):
+                    env_in[name] = str(path)
+                elif name in ("VALIDOUT", "REJECT", "REPORTERR", "CALCOUT", "STMT", "SUMMARY"):
+                    env_out[name] = str(path)
+
+        input_path = env_in.get(list(env_in.keys())[0], "")
+        output_path = env_out.get(list(env_out.keys())[0], str(self.root_dir / "data" / "work" / f"{step.step_name.lower()}_out.bin"))
+
+        run = runner.run(build.artifact_path, input_path, output_path)
+        self.results[step.step_name] = {
+            "status": "OK" if run.success else "ERROR",
+            "rc": 0 if run.success else 12
+        }
+        return 0 if run.success else 12
+
+    def _check_cond(self, cond: CondParam) -> bool:
+        if cond.step_name:
+            target = self.step_rcs.get(cond.step_name, 0)
+            return not COND_OPS.get(cond.operator, lambda x, y: False)(target, cond.code)
+        return True
+
+    def _run_sort(self, step: JobStep) -> int:
+        import subprocess
+        infile = None
+        outfile = None
+        for dd in step.dd_entries:
+            n = dd.dd_name.upper()
+            if n == "SORTIN" and dd.dsn:
+                infile = self._resolve_path(dd.dsn)
+            elif n == "SORTOUT" and dd.dsn:
+                outfile = self._resolve_path(dd.dsn)
+
+        if not infile or not outfile:
+            return 12
+
+        if infile.exists():
+            lines = sorted(infile.read_text().splitlines())
+            outfile.write_text("\n".join(lines))
+        self.results[step.step_name] = {"status": "OK", "rc": 0}
+        return 0
+
+    def _resolve_path(self, dsn: str) -> Path:
+        import re
+        dsn = re.sub(r"\(\+?\d+\)", "", dsn).strip(".")
+        return (self.root_dir / dsn.lstrip("/")).resolve()

jcl/parser.py

@@ -0,0 +1,118 @@
+"""JCL Parser — Phase 2. Parses JCL scripts into structured Job objects."""
+import re
+from dataclasses import dataclass, field
+from typing import Optional
+
+
+@dataclass
+class DDEntry:
+    dd_name: str
+    dsn: Optional[str] = None
+    disp: Optional[str] = None
+    sysout: Optional[str] = None
+    inline_data: list[str] = field(default_factory=list)
+
+
+@dataclass
+class CondParam:
+    code: int
+    operator: str   # EQ, NE, GT, GE, LT, LE
+    step_name: Optional[str] = None
+
+
+@dataclass
+class JobStep:
+    step_name: str
+    program: str
+    dd_entries: list[DDEntry] = field(default_factory=list)
+    cond: Optional[CondParam] = None
+    parm: Optional[str] = None
+
+
+@dataclass
+class Job:
+    job_name: str
+    steps: list[JobStep] = field(default_factory=list)
+
+
+COND_OPS = {
+    "EQ": lambda rc, code: rc == code, "NE": lambda rc, code: rc != code,
+    "GT": lambda rc, code: rc > code,  "GE": lambda rc, code: rc >= code,
+    "LT": lambda rc, code: rc < code,  "LE": lambda rc, code: rc <= code,
+}
+
+
+def parse_jcl(filepath: str) -> Optional[Job]:
+    """Parse a JCL file into a Job object."""
+    with open(filepath, "r", encoding="utf-8") as f:
+        lines = _merge_continuations(f.readlines())
+
+    job = None
+    current_step: Optional[JobStep] = None
+    in_sysin = False
+    sysin_lines: list[str] = []
+
+    for line in lines:
+        stripped = line.strip()
+        if stripped.startswith("//*") or not stripped:
+            continue
+
+        if in_sysin:
+            if stripped == "/*":
+                if current_step and current_step.dd_entries:
+                    current_step.dd_entries[-1].inline_data = sysin_lines
+                sysin_lines = []; in_sysin = False
+            else:
+                sysin_lines.append(stripped)
+            continue
+
+        if not stripped.startswith("//"):
+            continue
+
+        content = stripped[2:].strip()
+
+        # JOB
+        if re.search(r"\bJOB\b", content, re.IGNORECASE):
+            job = Job(job_name=content.split()[0])
+            continue
+
+        # EXEC
+        m = re.match(r"(\w+)\s+EXEC\s+(?:PGM=)?(\w+)", content, re.IGNORECASE)
+        if m:
+            step_name, program = m.group(1), m.group(2)
+            cond = None
+            cm = re.search(r"COND=\s*\(\s*(\d+)\s*,\s*(\w+)", content, re.IGNORECASE)
+            if cm:
+                cond = CondParam(code=int(cm.group(1)), operator=cm.group(2).upper())
+            current_step = JobStep(step_name=step_name, program=program, cond=cond)
+            if job:
+                job.steps.append(current_step)
+            continue
+
+        # DD
+        m = re.match(r"(\w+)\s+DD\s*(.*)", content, re.IGNORECASE)
+        if m and current_step is not None:
+            dd = DDEntry(dd_name=m.group(1))
+            params = m.group(2)
+            dm = re.search(r"DSN=\s*([^\s,]+)", params, re.IGNORECASE)
+            if dm:
+                dd.dsn = dm.group(1)
+            if dd.dd_name.upper() == "SYSIN" and "*" in params:
+                in_sysin = True
+            current_step.dd_entries.append(dd)
+            continue
+
+    return job
+
+
+def _merge_continuations(lines: list[str]) -> list[str]:
+    merged, buf = [], ""
+    for line in lines:
+        s = line.rstrip("\n\r")
+        buf = (buf + s) if buf else s
+        if s.rstrip().endswith(",") and not s.strip().startswith("//*"):
+            continue
+        merged.append(buf); buf = ""
+    if buf:
+        merged.append(buf)
+    return merged

orchestrator.py

@@ -10,6 +10,7 @@ from runners.cobol_runner import CobolRunner
 from runners.data_writer import DataWriter
 from agents.agent1_parser import Agent1Parser
 from agents.agent2_data import Agent2Data
+from agents.agent3_diagnostic import Agent3Diagnostic
 from agents.llm import LLMClient
 from comparator.aligner import align_records
 from comparator.field_compare import compare_field
@@ -31,6 +32,9 @@ def run_pipeline(cfg: Config, cpath: str, cbl: str, java: str, map_path: str) ->
         llm = LLMClient(model=cfg.llm_model, timeout=cfg.llm_timeout, cache_dir=cfg.llm_cache_dir)
         tree = Agent1Parser(llm).parse(text)
         vr.llm_cost += 0.002
+        vr.debug["field_tree"] = [{"name":f.name,"level":f.level,"pic":f.pic,"usage":f.usage,
+            "offset":f.offset,"length":f.length,"redefines":f.redefines}
+            for f in tree.flatten().values()]
         if not tree.fields:
             return _done(vr, t0, "BLOCKED", 2)
         if vr.llm_cost > cfg.max_llm_cost:
@@ -38,6 +42,8 @@ def run_pipeline(cfg: Config, cpath: str, cbl: str, java: str, map_path: str) ->
 
         suite = Agent2Data(llm).design(tree, cfg.coverage_default, cfg.runner_mode == "spark")
         vr.llm_cost += 0.002
+        vr.debug["test_cases"] = [{"id":tc.id,"fields":tc.fields,"targets":tc.coverage_targets} for tc in suite.test_cases]
+        vr.debug["spark_config"] = {"records":suite.spark_config.num_records} if suite.has_spark else None
 
         bundle = TestDataBundle(base_path=Path("test-data-bundle"))
         bundle.ensure_dirs()
@@ -51,6 +57,7 @@ def run_pipeline(cfg: Config, cpath: str, cbl: str, java: str, map_path: str) ->
 
         cob = CobolRunner()
         build = cob.compile(cbl, cfg.dialect)
+        vr.debug["cobol_build"] = {"ok": build.success, "log": build.log[-300:]}
         if not build.success:
             return _done(vr, t0, "BLOCKED", 2)
         co = Path("cobol_out.bin")
@@ -61,6 +68,7 @@ def run_pipeline(cfg: Config, cpath: str, cbl: str, java: str, map_path: str) ->
             return _done(vr, t0, "BLOCKED", 2)
         runner: Runner = SparkJavaRunner(cfg.spark_master) if cfg.runner_mode == "spark" else NativeJavaRunner()
         jb = runner.compile(java)
+        vr.debug["java_build"] = {"ok": jb.success, "log": jb.log[-300:]}
         if not jb.success:
             return _done(vr, t0, "BLOCKED", 2)
         inp = str(bundle.spark_input_dir() if cfg.runner_mode == "spark" else bundle.native_input())
@@ -95,6 +103,14 @@ def run_pipeline(cfg: Config, cpath: str, cbl: str, java: str, map_path: str) ->
         vr.status = "PASS" if m == 0 else "MISMATCH"
         vr.exit_code = 0 if m == 0 else 1
 
+        diag = Agent3Diagnostic(llm)
+        for fr in frs:
+            if fr.status in ("MISMATCH", "NOT_SET", "NPE"):
+                try:
+                    fr.suggestion = diag.analyze(fr) or ""
+                except:
+                    pass
+
         rd = Path(f"reports/{vr.program}") / vr.timestamp
         rd.mkdir(parents=True, exist_ok=True)
         g = ReportGenerator()

tasks/6aaa7e43.json

@@ -0,0 +1 @@
+{"id": "6aaa7e43", "status": "done", "copybook": "uploads\\6aaa7e43\\copybook.cpy", "cobol_src": "uploads\\6aaa7e43\\program.cbl", "java_src": "uploads\\6aaa7e43\\java", "mapping": "uploads\\6aaa7e43\\mapping.yaml", "runner": "native", "created": "2026-05-25T09:19:06.763502", "fields": [], "result": {"program": "java", "status": "ERROR", "matched": 0, "mismatched": 0, "duration": 34.83706521987915, "runner": "native"}}

tests/fixtures/simple.cbl

@@ -7,7 +7,10 @@
           05 BR-STATUS PIC X.
           05 BR-DATE   PIC 9(8).
        PROCEDURE DIVISION.
-           DISPLAY BR-AMT.
+           MOVE 1500 TO BR-AMT.
-           DISPLAY BR-STATUS.
+           MOVE 'A' TO BR-STATUS.
-           DISPLAY BR-DATE.
+           MOVE 20260522 TO BR-DATE.
+           DISPLAY BR-AMT
+           DISPLAY BR-STATUS
+           DISPLAY BR-DATE
            STOP RUN.

tests/test_golden.py

@@ -0,0 +1,226 @@
+"""
+Golden tests — 对接真实 COBOL 信用卡月结系统
+数据源: D:\cobol-java\jcl-cobol-git\
+验证目标: 28 transactions → 20 valid + 8 rejected → 6 cards → ¥48,250.20 total
+"""
+import sys, os
+sys.path.insert(0, r"D:\cobol-java\v3-gstack-code-gen")
+
+from pathlib import Path
+from data.field_tree import Field, FieldTree
+from comparator.cobol_binary_reader import CobolBinaryReader
+from comparator.normalizer import Normalizer
+from comparator.field_compare import compare_field
+from comparator.aligner import align_records
+from preprocessor import CopybookPreprocessor
+
+GOLDEN = Path(r"D:\cobol-java\jcl-cobol-git")
+
+
+# ── Test 1: TXCPY COPYBOOK 结构验证 ──
+def test_txcpy_field_count():
+    """TXCPY COPYBOOK defines 6 top-level fields."""
+    txcpy = FieldTree(fields=[
+        Field(name="TX-CARD-NO",  level=5, pic="9(16)",      usage="DISPLAY", offset=0,  length=16),
+        Field(name="TX-DATE",     level=5, pic="9(8)",       usage="DISPLAY", offset=16, length=8),
+        Field(name="TX-TYPE",     level=5, pic="X",          usage="DISPLAY", offset=24, length=1),
+        Field(name="TX-AMOUNT",   level=5, pic="S9(9)V99",  usage="DISPLAY", offset=25, length=11),
+        Field(name="TX-CURRENCY", level=5, pic="X(3)",       usage="DISPLAY", offset=36, length=3),
+        Field(name="TX-MERCHANT", level=5, pic="X(20)",      usage="DISPLAY", offset=39, length=20),
+    ], copybook_name="TXCPY")
+    flat = txcpy.flatten()
+    assert len(flat) == 6, f"Expected 6 fields, got {len(flat)}"
+    assert flat["TX-CARD-NO"].length == 16
+    assert flat["TX-AMOUNT"].length == 11
+
+
+# ── Test 2: 二进制文件读取 ──
+def test_read_transactions():
+    """读取交易数据，验证第一条记录结构正确。"""
+    txcpy = FieldTree(fields=[
+        Field(name="TX-CARD-NO",  level=5, pic="9(16)",      usage="DISPLAY", offset=0,  length=16),
+        Field(name="TX-DATE",     level=5, pic="9(8)",       usage="DISPLAY", offset=16, length=8),
+        Field(name="TX-TYPE",     level=5, pic="X",          usage="DISPLAY", offset=24, length=1),
+        Field(name="TX-AMOUNT",   level=5, pic="S9(9)V99",  usage="DISPLAY", offset=25, length=12),
+        Field(name="TX-CURRENCY", level=5, pic="X(3)",       usage="DISPLAY", offset=37, length=3),
+        Field(name="TX-MERCHANT", level=5, pic="X(20)",      usage="DISPLAY", offset=40, length=20),
+    ], copybook_name="TXCPY")
+
+    reader = CobolBinaryReader()
+    records = reader.read(str(GOLDEN / "data/input/transactions.dat"), txcpy)
+
+    # 文件 2044 bytes, record size ≈ 73, 大约 28 条记录
+    assert len(records) >= 28, f"Expected >=28, got {len(records)}"
+
+    r0 = records[0]
+    assert r0["TX-CARD-NO"] == "6222021234567800"
+    assert r0["TX-DATE"] == "20260501"
+    assert r0["TX-TYPE"] == "P"
+
+
+# ── Test 3: COMP-3 RATE 解析 ──
+def test_comp3_rate():
+    """验证利率表 COMP-3 编码。"""
+    n = Normalizer()
+    data = (GOLDEN / "data/input/rate.dat").read_bytes()
+    
+    assert len(data) == 24, "2 records × 12 bytes"
+    
+    # Record 1: Cash rate
+    assert n.normalize_comp3(data[1:4]) == "5"       # 0.0005
+    assert chr(data[0]) == "C"
+    
+    # Record 2: Overdue rate
+    assert n.normalize_comp3(data[13:16]) == "500"    # 0.0500
+    assert chr(data[12]) == "O"
+
+
+# ── Test 4: 管道输出一致性 ──
+def test_pipeline_counts():
+    """验证 28→20+8→6 的管道计数。"""
+    validated = len((GOLDEN / "data/work/validated_tx.dat").read_text().splitlines())
+    rejected = len((GOLDEN / "data/output/rejected_tx.dat").read_text().splitlines())
+    
+    assert validated == 20, f"Expected 20 valid, got {validated}"
+    assert rejected == 8, f"Expected 8 rejected, got {rejected}"
+
+
+def test_error_report_coverage():
+    """验证全部 7 条校验规则被触发。"""
+    errors = (GOLDEN / "data/output/error_report.dat").read_text()
+    
+    expected_errors = ["INVALID-CARD", "FROZEN-CARD", "INVALID-MERCHANT",
+                       "INVALID-AMOUNT", "INVALID-REFUND", "OUT-OF-MONTH", "MEMBER-NOT-FOUND"]
+    for e in expected_errors:
+        assert e in errors, f"Missing error: {e}"
+
+
+def test_grand_total():
+    """验证全局合计金额。"""
+    summary = (GOLDEN / "data/output/summary_report.dat").read_text()
+    assert "AMOUNT:        48250.20" in summary
+    assert "INTEREST:          300.00" in summary
+    assert "FEE:          800.00" in summary
+    assert "CARDS:00006" in summary
+
+
+# ── Test 5: COPY REPLACING 展开 ──
+def test_copy_replacing_datesub():
+    """验证 DATESUB COPYBOOK 的 REPLACING 展开。"""
+    pp = CopybookPreprocessor(paths=[str(GOLDEN / "copybooks")])
+    source = "        COPY DATESUB REPLACING ==:TAG:== BY ==WS-RUN==."
+    result = pp.expand(source)
+    assert "WS-RUN" in result or "DATESUB" in result
+
+
+# ── Test 6: 比对引擎集成 ──
+def test_compare_pipeline_output():
+    """验证比对引擎可以处理同源数据的 self-compare（应全部 PASS）。"""
+    reader = CobolBinaryReader()
+    txcpy = FieldTree(fields=[
+        Field(name="TX-CARD-NO",  level=5, pic="9(16)",      usage="DISPLAY", offset=0,  length=16),
+        Field(name="TX-DATE",     level=5, pic="9(8)",       usage="DISPLAY", offset=16, length=8),
+        Field(name="TX-TYPE",     level=5, pic="X",          usage="DISPLAY", offset=24, length=1),
+        Field(name="TX-AMOUNT",   level=5, pic="S9(9)V99",  usage="DISPLAY", offset=25, length=11),
+        Field(name="TX-CURRENCY", level=5, pic="X(3)",       usage="DISPLAY", offset=36, length=3),
+        Field(name="TX-MERCHANT", level=5, pic="X(20)",      usage="DISPLAY", offset=39, length=20),
+    ], copybook_name="TXCPY")
+    
+    records = reader.read(str(GOLDEN / "data/input/transactions.dat"), txcpy)
+    
+    # Self-compare: 同源数据应对齐后全 PASS
+    aligned = align_records(records, records, key_field="TX-CARD-NO")
+    assert len(aligned) == len(records), f"Alignment lost records: {len(aligned)} vs {len(records)}"
+    
+    # 每个字段 self-compare 应 PASS
+    for c, j, _ in aligned:
+        for key in ["TX-TYPE", "TX-CURRENCY", "TX-MERCHANT"]:
+            fr = compare_field(key, str(c.get(key, "")), str(j.get(key, "")), "string")
+            assert fr.status == "PASS", f"Self-compare failed: {key}"
+
+
+# ── JCL Tests ──
+
+def test_jcl_parse():
+    """验证 JCL 解析器正确解析 CREDIT25.jcl"""
+    import sys
+    sys.path.insert(0, str(GOLDEN.parent / "v3-gstack-code-gen"))
+    from jcl.parser import parse_jcl
+
+    jcl_path = str(GOLDEN / "jcl" / "CREDIT25.jcl")
+    job = parse_jcl(jcl_path)
+
+    assert job is not None, "JCL parse returned None"
+    assert job.job_name == "CREDIT25", f"Expected CREDIT25, got {job.job_name}"
+    assert len(job.steps) == 4, f"Expected 4 steps, got {len(job.steps)}"
+
+    # Step 1: SORT
+    assert job.steps[0].step_name == "STEP1"
+    assert job.steps[0].program == "SORT"
+    assert len(job.steps[0].dd_entries) == 3  # SORTIN, SORTOUT, SYSIN
+    assert any(dd.dd_name.upper() == "SORTIN" for dd in job.steps[0].dd_entries)
+    assert any(dd.dd_name.upper() == "SORTOUT" for dd in job.steps[0].dd_entries)
+
+    # Step 2: CRDVAL with COND
+    assert job.steps[1].step_name == "STEP2"
+    assert job.steps[1].program == "CRDVAL"
+    assert job.steps[1].cond is not None
+    assert job.steps[1].cond.code == 0
+    assert job.steps[1].cond.operator == "NE"
+    assert len(job.steps[1].dd_entries) == 6  # TRANSIN, MEMBER, VALIDOUT, REJECT, REPORTERR, SYSOUT
+
+    # Step 3: CRDCALC with COND  
+    assert job.steps[2].step_name == "STEP3"
+    assert job.steps[2].program == "CRDCALC"
+    assert job.steps[2].cond.code == 0
+
+    # Step 4: CRDRPT with COND
+    assert job.steps[3].step_name == "STEP4"
+    assert job.steps[3].program == "CRDRPT"
+    assert job.steps[3].cond.code == 0
+
+
+def test_jcl_dd_mapping():
+    """验证 JCL DD 语句正确提取"""
+    import sys
+    sys.path.insert(0, str(GOLDEN.parent / "v3-gstack-code-gen"))
+    from jcl.parser import parse_jcl
+
+    job = parse_jcl(str(GOLDEN / "jcl" / "CREDIT25.jcl"))
+
+    # Check DD names for CRDVAL step
+    crdval_dds = {dd.dd_name.upper() for dd in job.steps[1].dd_entries}
+    assert crdval_dds >= {"TRANSIN", "MEMBER", "VALIDOUT", "REJECT", "REPORTERR"}, \
+        f"Missing DDs: {crdval_dds}"
+
+    # Check DD names for CRDCALC step
+    crdcalc_dds = {dd.dd_name.upper() for dd in job.steps[2].dd_entries}
+    assert crdcalc_dds >= {"VALIDIN", "RATE", "CALCOUT"}, \
+        f"Missing DDs: {crdcalc_dds}"
+
+    # Check DD names for CRDRPT step
+    crdrpt_dds = {dd.dd_name.upper() for dd in job.steps[3].dd_entries}
+    assert crdrpt_dds >= {"BILLING", "STMT", "SUMMARY"}, \
+        f"Missing DDs: {crdrpt_dds}"
+
+
+def test_jcl_job_info():
+    """验证 JCL Job 基本信息"""
+    import sys
+    sys.path.insert(0, str(GOLDEN.parent / "v3-gstack-code-gen"))
+    from jcl.parser import parse_jcl
+
+    job = parse_jcl(str(GOLDEN / "jcl" / "CREDIT25.jcl"))
+
+    assert job.job_name == "CREDIT25"
+    assert len(job.steps) == 4
+
+    programs = [s.program for s in job.steps]
+    assert programs == ["SORT", "CRDVAL", "CRDCALC", "CRDRPT"], \
+        f"Unexpected program order: {programs}"
+
+    # Verify all COND are (0,NE) — run step if previous step succeeded
+    for i in [1, 2, 3]:
+        assert job.steps[i].cond is not None, f"Step {i+1} missing COND"
+        assert job.steps[i].cond.code == 0
+        assert job.steps[i].cond.operator == "NE"

uploads/6aaa7e43/copybook.cpy

@@ -0,0 +1,4 @@
+01 BILL-RECORD.
+   05 BR-AMT    PIC S9(7)V99 COMP-3.
+   05 BR-STATUS PIC X.
+   05 BR-DATE   PIC 9(8).

uploads/6aaa7e43/java

@@ -0,0 +1 @@
+<project><modelVersion>4.0.0</modelVersion><groupId>test</groupId><artifactId>test</artifactId><version>1.0</version><properties><maven.compiler.source>17</maven.compiler.source><maven.compiler.target>17</maven.compiler.target></properties><dependencies><dependency><groupId>com.fasterxml.jackson.core</groupId><artifactId>jackson-databind</artifactId><version>2.17.0</version></dependency></dependencies></project>

uploads/6aaa7e43/mapping.yaml

@@ -0,0 +1,13 @@
+program: SIMPLE
+field_mapping:
+  - cobol_field: BR-AMT
+    java_field: billAmount
+    type: decimal
+    precision: 2
+  - cobol_field: BR-STATUS
+    java_field: statusCode
+    type: string
+  - cobol_field: BR-DATE
+    java_field: billDate
+    type: date
+    format: YYYYMMDD

uploads/6aaa7e43/program.cbl

@@ -0,0 +1,13 @@
+       IDENTIFICATION DIVISION.
+       PROGRAM-ID. SIMPLE.
+       DATA DIVISION.
+       WORKING-STORAGE SECTION.
+       01 BILL-RECORD.
+          05 BR-AMT    PIC S9(7)V99 COMP-3.
+          05 BR-STATUS PIC X.
+          05 BR-DATE   PIC 9(8).
+       PROCEDURE DIVISION.
+           DISPLAY BR-AMT.
+           DISPLAY BR-STATUS.
+           DISPLAY BR-DATE.
+           STOP RUN.

web/api.py

@@ -1,80 +1,89 @@
 """Web API layer — wraps orchestrator with 202+ polling pattern."""
-import uuid, json, shutil, sys, os
+import uuid, json, sys, os
 from pathlib import Path
 from datetime import datetime
 from fastapi import FastAPI, UploadFile, File, Form, HTTPException
 from fastapi.responses import HTMLResponse, JSONResponse
 from fastapi.staticfiles import StaticFiles
-from fastapi.templating import Jinja2Templates
-from starlette.requests import Request
 
 sys.path.insert(0, str(Path(__file__).parent.parent))
 from config import Config
 from orchestrator import run_pipeline
 
-app = FastAPI(title="COBOL→Java Verify")
+app = FastAPI(title="COBOL->Java Verify")
-app.mount("/static", StaticFiles(directory=str(Path(__file__).parent / "static")), name="static")
+BASE = Path(__file__).parent
-templates = Jinja2Templates(directory=str(Path(__file__).parent / "templates"))
+app.mount("/static", StaticFiles(directory=str(BASE / "static")), name="static")
 
-TASKS_DIR = Path("tasks")
+TASKS_DIR = Path("tasks"); TASKS_DIR.mkdir(exist_ok=True)
-TASKS_DIR.mkdir(exist_ok=True)
+UPLOAD_DIR = Path("uploads"); UPLOAD_DIR.mkdir(exist_ok=True)
-UPLOAD_DIR = Path("uploads")
+MAX_SIZE = 10 * 1024 * 1024
-UPLOAD_DIR.mkdir(exist_ok=True)
-MAX_SIZE = 10 * 1024 * 1024  # 10MB
 
 
 @app.get("/", response_class=HTMLResponse)
-async def index(request: Request):
+async def index():
-    return templates.TemplateResponse("upload.html", {"request": request})
+    return (BASE / "templates" / "upload.html").read_text(encoding="utf-8")
 
 
 @app.post("/verify")
 async def verify(
-    copybook: UploadFile = File(...),
+    copybook: UploadFile = File(...), cobol_src: UploadFile = File(...),
-    cobol_src: UploadFile = File(...),
+    java_src: UploadFile = File(...), mapping: UploadFile = File(...),
-    java_src: UploadFile = File(...),
-    mapping: UploadFile = File(...),
     runner: str = Form("native"),
 ):
     task_id = str(uuid.uuid4())[:8]
-    task_dir = UPLOAD_DIR / task_id
+    task_dir = UPLOAD_DIR / task_id; task_dir.mkdir(parents=True, exist_ok=True)
-    task_dir.mkdir(parents=True, exist_ok=True)
+    for f, name in [(copybook,"copybook.cpy"),(cobol_src,"program.cbl"),
-
+                     (java_src,"java"),(mapping,"mapping.yaml")]:
-    for f, name in [(copybook, "copybook.cpy"), (cobol_src, "program.cbl"),
-                     (java_src, "java"), (mapping, "mapping.yaml")]:
         content = await f.read()
         if len(content) > MAX_SIZE:
-            raise HTTPException(413, f"{f.filename} exceeds 10MB limit")
+            raise HTTPException(413, f"{f.filename} exceeds 10MB")
-        dest = task_dir / name
+        (task_dir / name).write_bytes(content)
-        dest.write_bytes(content)
+    (TASKS_DIR / f"{task_id}.json").write_text(json.dumps({
-
+        "id":task_id,"status":"queued","copybook":str(task_dir/"copybook.cpy"),
-    task_file = TASKS_DIR / f"{task_id}.json"
+        "cobol_src":str(task_dir/"program.cbl"),"java_src":str(task_dir/"java"),
-    task_file.write_text(json.dumps({
+        "mapping":str(task_dir/"mapping.yaml"),"runner":runner,
-        "id": task_id, "status": "queued",
+        "created":datetime.now().isoformat()}))
-        "copybook": str(task_dir / "copybook.cpy"),
+    return JSONResponse({"task_id":task_id,"status":"queued"}, status_code=202)
-        "cobol_src": str(task_dir / "program.cbl"),
-        "java_src": str(task_dir / "java"),
-        "mapping": str(task_dir / "mapping.yaml"),
-        "runner": runner, "created": datetime.now().isoformat()
-    }))
-
-    return JSONResponse({"task_id": task_id, "status": "queued"}, status_code=202)
 
 
 @app.get("/status/{task_id}")
 async def status(task_id: str):
     tf = TASKS_DIR / f"{task_id}.json"
-    if not tf.exists():
+    if not tf.exists(): raise HTTPException(404, "task not found")
-        raise HTTPException(404, "task not found")
     data = json.loads(tf.read_text())
-    return JSONResponse({"task_id": task_id, "status": data.get("status", "unknown"),
+    return JSONResponse({"task_id":task_id,"status":data.get("status","unknown"),
-                         "result": data.get("result")})
+                         "result":data.get("result"),"fields":data.get("fields",[])})
+
+
+@app.get("/fields/{task_id}")
+async def fields(task_id: str):
+    tf = TASKS_DIR / f"{task_id}.json"
+    if not tf.exists(): raise HTTPException(404, "task not found")
+    data = json.loads(tf.read_text())
+    return JSONResponse({"task_id":task_id,"fields":data.get("fields",[]),
+                         "debug":data.get("debug",{}),
+                         "build_log":data.get("build_log","")})
 
 
 @app.get("/result/{task_id}", response_class=HTMLResponse)
-async def result(request: Request, task_id: str):
+async def result(task_id: str):
     tf = TASKS_DIR / f"{task_id}.json"
-    if not tf.exists():
+    if not tf.exists(): raise HTTPException(404, "task not found")
-        raise HTTPException(404, "task not found")
     data = json.loads(tf.read_text())
-    return templates.TemplateResponse("result.html", {"request": request, "task": data})
+    html = (BASE / "templates" / "result.html").read_text(encoding="utf-8")
+    html = html.replace("{{ task.id }}", data.get("id", task_id))
+    if data.get("status") == "done" and data.get("result"):
+        r = data["result"]
+        html = html.replace("{{ task.status }}", "done")
+        html = html.replace("{{ task.result.status }}", r.get("status",""))
+        html = html.replace("{{ task.result.program }}", r.get("program",""))
+        html = html.replace("{{ task.result.matched }}", str(r.get("matched",0)))
+        html = html.replace("{{ task.result.mismatched }}", str(r.get("mismatched",0)))
+        html = html.replace("{{ task.result.runner }}", r.get("runner",""))
+        html = html.replace("{{ task.result.duration }}", str(r.get("duration",0)))
+    elif data.get("status") == "error":
+        html = html.replace("{{ task.status }}", "error")
+        html = html.replace("{{ task.result.status }}", data.get("result",""))
+    else:
+        html = html.replace("{{ task.status }}", data.get("status","queued"))
+    return HTMLResponse(html)

web/static/script.js

@@ -1,25 +1,36 @@
 document.getElementById("verify-form").addEventListener("submit", async (e) => {
   e.preventDefault();
-  const btn = e.target.querySelector("button");
+  const btn = e.target.querySelector("button[type=submit]");
   btn.disabled = true;
-  btn.textContent = "Submitting...";
+  btn.textContent = "$ uploading...";
-  document.getElementById("status").textContent = "";
+  document.getElementById("status-area").innerHTML = "";
-  document.getElementById("result").innerHTML = "";
 
   const fd = new FormData(e.target);
   try {
     const r = await fetch("/verify", { method: "POST", body: fd });
     const d = await r.json();
     if (r.ok) {
-      document.getElementById("status").textContent = "Task " + d.task_id + " queued";
+      document.getElementById("status-area").innerHTML = `
-      document.getElementById("result").innerHTML =
+        <div class="status-card pending">
-        '<p><a href="/result/' + d.task_id + '">View result →</a></p>';
+          <div class="title">&#9679; Queued</div>
+          Task <code>${d.task_id}</code> submitted. Worker processing.
+          <div class="matrix"><dt>Runner</dt><dd>${fd.get("runner")}</dd></div>
+          <a class="result-link" href="/result/${d.task_id}">Open result page &rarr;</a>
+        </div>`;
     } else {
-      document.getElementById("status").textContent = "Error: " + (d.detail || "unknown");
+      document.getElementById("status-area").innerHTML = `
+        <div class="status-card error">
+          <div class="title">&#10007; Error</div>
+          ${d.detail || "Upload failed"}
+        </div>`;
     }
   } catch (err) {
-    document.getElementById("status").textContent = "Upload failed: " + err.message;
+    document.getElementById("status-area").innerHTML = `
+      <div class="status-card error">
+        <div class="title">&#10007; Network Error</div>
+        ${err.message}
+      </div>`;
   }
   btn.disabled = false;
-  btn.textContent = "Verify";
+  btn.textContent = "$ verify";
 });

web/static/style.css

@@ -1,11 +1,70 @@
-body { font-family: monospace; max-width: 900px; margin: 2rem auto; padding: 0 1rem; color: #333; }
+:root {
-h1 { font-size: 1.2rem; }
+  --bg: #0a0e14;
-label { display: block; margin: .75rem 0; }
+  --panel: #12171f;
-input, select { display: block; margin-top: .25rem; }
+  --border: #1f2937;
-button { margin-top: 1rem; padding: .5rem 1.5rem; cursor: pointer; font-size: 1rem; }
+  --text: #b2becd;
-#status { margin-top: 1rem; font-weight: bold; }
+  --dim: #5c6e80;
-#result { margin-top: 1rem; }
+  --accent: #39bae6;
-pre { background: #f0f0f0; padding: 1rem; border-radius: 4px; }
+  --green: #7fd962;
-.pass { border-left: 4px solid #4caf50; }
+  --red: #f26d78;
-.fail { border-left: 4px solid #f44336; }
+  --yellow: #ffad66;
-.container a { display: inline-block; margin-top: 1rem; }
+  --font: "SF Mono", "Fira Code", "Cascadia Code", Consolas, monospace;
+}
+* { box-sizing: border-box; margin: 0; padding: 0; }
+body { font-family: var(--font); background: var(--bg); color: var(--text); line-height: 1.6; }
+body::before { content:""; position:fixed; top:0; left:0; width:100%; height:100%; background: radial-gradient(ellipse at 20% 20%, rgba(57,186,230,.03) 0%, transparent 50%), radial-gradient(ellipse at 80% 80%, rgba(127,217,98,.02) 0%, transparent 50%); pointer-events:none; z-index:0; }
+.container { max-width: 680px; margin: 0 auto; padding: 3rem 1.5rem; position:relative; z-index:1; }
+/* Header */
+header { margin-bottom: 2.5rem; border-bottom: 1px solid var(--border); padding-bottom: 1.5rem; }
+.badge { display:inline-block; padding:.15rem .6rem; background:rgba(57,186,230,.12); color:var(--accent); border-radius:3px; font-size:.7rem; text-transform:uppercase; letter-spacing:.08em; margin-bottom:.75rem; }
+h1 { font-size:1.3rem; font-weight:400; color:#e6edf3; line-height:1.3; }
+h1 span { color: var(--accent); }
+.tagline { font-size:.8rem; color:var(--dim); margin-top:.4rem; }
+/* Sections */
+section { background:var(--panel); border:1px solid var(--border); border-radius:8px; padding:1.5rem; margin-bottom:1rem; }
+section h2 { font-size:.8rem; color:var(--dim); text-transform:uppercase; letter-spacing:.06em; margin-bottom:1rem; font-weight:400; }
+/* Form */
+.form-grid { display:grid; grid-template-columns:1fr 1fr; gap:.75rem; }
+.form-grid .full { grid-column:1/-1; }
+label { font-size:.75rem; color:var(--dim); }
+label strong { display:block; color:var(--text); font-weight:400; margin-bottom:.25rem; font-size:.8rem; }
+input[type="file"], select { width:100%; padding:.55rem .7rem; background:var(--bg); border:1px solid var(--border); border-radius:6px; color:var(--text); font-family:inherit; font-size:.8rem; margin-top:.15rem; transition:border-color .2s; }
+input[type="file"]:hover, select:hover { border-color: var(--accent); }
+input[type="file"]::file-selector-button { background:var(--border); color:var(--text); border:none; padding:.35rem .8rem; border-radius:4px; margin-right:.5rem; cursor:pointer; font-family:inherit; font-size:.75rem; }
+select { cursor:pointer; background-image:url("data:image/svg+xml,%3Csvg xmlns='http://www.w3.org/2000/svg' width='12' height='12' fill='%235c6e80'%3E%3Cpath d='M6 8L2 4h8z'/%3E%3C/svg%3E"); background-repeat:no-repeat; background-position:right .5rem center; appearance:none; padding-right:2rem; }
+/* Buttons */
+.actions { display:flex; gap:.5rem; margin-top:1.25rem; }
+.btn { padding:.6rem 1.25rem; border-radius:6px; font-family:inherit; font-size:.8rem; cursor:pointer; border:none; transition: all .2s; }
+.btn-primary { background:var(--accent); color:var(--bg); }
+.btn-primary:hover { background:#4fc8f0; }
+.btn-primary:disabled { background:var(--border); color:var(--dim); cursor:not-allowed; }
+.btn-secondary { background:transparent; color:var(--dim); border:1px solid var(--border); }
+.btn-secondary:hover { color:var(--text); border-color:var(--dim); }
+/* Status & Results */
+#status-area { margin-top:1rem; }
+.status-card { background:var(--panel); border:1px solid var(--border); border-radius:8px; padding:1rem 1.5rem; font-size:.8rem; }
+.status-card.success { border-left:3px solid var(--green); }
+.status-card.pending { border-left:3px solid var(--yellow); }
+.status-card.error { border-left:3px solid var(--red); }
+.status-card .title { font-weight:600; margin-bottom:.25rem; }
+.status-card.success .title { color:var(--green); }
+.status-card.pending .title { color:var(--yellow); }
+.status-card.error .title { color:var(--red); }
+.matrix { display:grid; grid-template-columns:auto 1fr; gap:.3rem 1.5rem; margin:.75rem 0; font-size:.8rem; }
+.matrix dt { color:var(--dim); }
+.matrix dd { color:var(--text); }
+.divider { border-top:1px solid var(--border); margin:1rem 0; }
+.result-link { display:inline-block; margin-top:.5rem; color:var(--accent); text-decoration:none; font-size:.8rem; }
+.result-link:hover { text-decoration:underline; }
+/* Footer */
+footer { margin-top:3rem; padding-top:1rem; border-top:1px solid var(--border); font-size:.7rem; color:var(--dim); display:flex; justify-content:space-between; }
+footer a { color:var(--dim); text-decoration:none; }
+footer a:hover { color:var(--text); }
+/* Field table */
+table { width:100%; border-collapse:collapse; font-size:.8rem; }
+th { text-align:left; color:var(--dim); font-weight:400; padding:.5rem .75rem; border-bottom:1px solid var(--border); }
+td { padding:.45rem .75rem; border-bottom:1px solid rgba(31,41,55,.5); }
+tr.pass td:first-child { border-left:3px solid var(--green); }
+tr.tolerated td:first-child { border-left:3px solid var(--yellow); }
+tr.fail td:first-child { border-left:3px solid var(--red); }
+pre { background:var(--bg); border:1px solid var(--border); border-radius:6px; padding:1rem; font-family:inherit; font-size:.8rem; overflow-x:auto; margin-top:.5rem; }

web/templates/result.html

@@ -1,27 +1,99 @@
 <!DOCTYPE html>
-<html><head><meta charset="utf-8"><title>Result - {{ task.id }}</title>
+<html lang="en">
-<link rel="stylesheet" href="/static/style.css"></head><body>
+<head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
+<title>Result — {{ task.id }}</title>
+<link rel="stylesheet" href="/static/style.css">
+</head>
+<body>
 <div class="container">
-<h1>Verification Result</h1>
+
-{% if task.status == "done" and task.result %}
+  <header>
-<pre class="pass">Status: {{ task.result.status }}
+    <div class="badge">Verification Result</div>
-Program: {{ task.result.program }}
+    <h1>{{ task.id }}</h1>
-Matched: {{ task.result.matched }} | Mismatched: {{ task.result.mismatched }}
+    <div class="tagline">Status: <span class="poll-status">loading...</span></div>
-Runner: {{ task.result.runner }} | Duration: {{ task.result.duration }}s</pre>
+  </header>
-{% elif task.status == "error" %}
+
-<pre class="fail">{{ task.result }}</pre>
+  <section>
-{% else %}
+    <h2>Summary</h2>
-<div id="poll-status">Status: {{ task.status }} — polling...</div>
+    <div class="matrix">
+      <dt>Status</dt><dd>{{ task.result.status }}</dd>
+      <dt>Program</dt><dd>{{ task.result.program }}</dd>
+      <dt>Matched</dt><dd>{{ task.result.matched }}</dd>
+      <dt>Mismatched</dt><dd>{{ task.result.mismatched }}</dd>
+      <dt>Runner</dt><dd>{{ task.result.runner }}</dd>
+      <dt>Duration</dt><dd>{{ task.result.duration }}s</dd>
+    </div>
+  </section>
+
+  <section>
+    <h2>Field Results</h2>
+    <div id="fields-table"></div>
+  </section>
+
+  <section>
+    <h2>Pipeline Details</h2>
+    <div id="debug-section"></div>
+  </section>
+
+  <div class="divider"></div>
+  <a href="/" class="btn btn-secondary">← New Verification</a>
+
+  <footer>
+    <span>verify-cli v0.2.0</span>
+    <span><a href="#">Docs</a> &middot; <a href="#">Architecture</a></span>
+  </footer>
+
+</div>
 <script>
 const id = "{{ task.id }}";
 setInterval(async () => {
   const r = await fetch("/status/" + id);
   const d = await r.json();
-  document.getElementById("poll-status").textContent = "Status: " + d.status;
+  document.querySelector(".poll-status").textContent = d.status;
   if (d.status === "done" || d.status === "error") location.reload();
 }, 3000);
+
+// Load per-field results + debug
+(async () => {
+  const r = await fetch("/fields/" + id);
+  const d = await r.json();
+  if (d.fields && d.fields.length) {
+    const rows = d.fields.map(f => {
+      const cls = f.status === "PASS" ? "pass" : f.status === "TOLERATED" ? "tolerated" : "fail";
+      return `<tr class="${cls}"><td>${f.name}</td><td>${f.status}</td><td>${f.cobol||""}</td><td>${f.java||""}</td><td>${f.suggestion||""}</td></tr>`;
+    }).join("");
+    document.getElementById("fields-table").innerHTML =
+      `<table><tr><th>Field</th><th>Status</th><th>COBOL</th><th>Java</th><th>Suggestion</th></tr>${rows}</table>`;
+  }
+  // Render debug info
+  const dbg = d.debug || {};
+  let html = "";
+  if (dbg.field_tree) {
+    const treeRows = dbg.field_tree.map(f =>
+      `<tr><td>L${f.level}</td><td>${f.name}</td><td>${f.pic}</td><td>${f.usage}</td><td>${f.offset}</td><td>${f.length}</td></tr>`).join("");
+    html += `<h3 style="color:var(--dim);font-size:.75rem;letter-spacing:.05em;margin:1rem 0 .5rem">COPYBOOK FieldTree</h3>
+      <table><tr><th>Lv</th><th>Name</th><th>PIC</th><th>Usage</th><th>Off</th><th>Len</th></tr>${treeRows}</table>`;
+  }
+  if (dbg.test_cases) {
+    const tcList = dbg.test_cases.map(tc =>
+      `<tr><td>${tc.id}</td><td>${JSON.stringify(tc.fields)}</td><td>${(tc.targets||[]).join(", ")}</td></tr>`).join("");
+    html += `<h3 style="color:var(--dim);font-size:.75rem;letter-spacing:.05em;margin:1rem 0 .5rem">Test Data (${dbg.test_cases.length} cases)</h3>
+      <table><tr><th>ID</th><th>Fields</th><th>Coverage</th></tr>${tcList}</table>`;
+  }
+  if (dbg.spark_config) {
+    html += `<h3 style="color:var(--dim);font-size:.75rem;letter-spacing:.05em;margin:1rem 0 .5rem">Spark Config</h3>
+      <pre>${dbg.spark_config.records} records via key_varied replication</pre>`;
+  }
+  if (dbg.cobol_build) {
+    html += `<h3 style="color:var(--dim);font-size:.75rem;letter-spacing:.05em;margin:1rem 0 .5rem">COBOL Compile${dbg.cobol_build.ok?"":" (FAILED)"}</h3>
+      <pre>${dbg.cobol_build.log||"(no output)"}</pre>`;
+  }
+  if (dbg.java_build) {
+    html += `<h3 style="color:var(--dim);font-size:.75rem;letter-spacing:.05em;margin:1rem 0 .5rem">Java Build${dbg.java_build.ok?"":" (FAILED)"}</h3>
+      <pre>${dbg.java_build.log||"(no output)"}</pre>`;
+  }
+  if (html) document.getElementById("debug-section").innerHTML = html;
+})();
 </script>
-{% endif %}
+</body>
-<a href="/">← New verification</a>
+</html>
-</div>
-</body></html>

web/templates/upload.html

@@ -1,18 +1,43 @@
 <!DOCTYPE html>
-<html><head><meta charset="utf-8"><title>COBOL→Java Verify</title>
+<html lang="en">
-<link rel="stylesheet" href="/static/style.css"></head><body>
+<head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
+<title>COBOL → Java Migration Verification</title>
+<link rel="stylesheet" href="/static/style.css">
+</head>
+<body>
 <div class="container">
-<h1>COBOL→Java Migration Verification</h1>
+
-<form id="verify-form" enctype="multipart/form-data">
+  <header>
-<label>COPYBOOK:<input type="file" name="copybook" accept=".cpy,.cbl,.copy" required></label>
+    <div class="badge">Developer Tool</div>
-<label>COBOL source:<input type="file" name="cobol_src" accept=".cbl" required></label>
+    <h1><span>verify</span> — COBOL to Java/Spark Migration</h1>
-<label>Java source (dir with pom.xml):<input type="file" name="java_src" webkitdirectory required></label>
+    <div class="tagline">Automated field-level verification for COBOL→Java migration pipelines</div>
-<label>Mapping YAML:<input type="file" name="mapping" accept=".yaml,.yml" required></label>
+  </header>
-<label>Runner:<select name="runner"><option value="native">Native Java</option><option value="spark">Spark Java</option></select></label>
+
-<button type="submit">Verify</button>
+  <section>
-</form>
+    <h2>Upload Sources</h2>
-<div id="status"></div>
+    <form id="verify-form" enctype="multipart/form-data">
-<div id="result"></div>
+      <div class="form-grid">
+        <label><strong>COPYBOOK</strong>.cpy / .cbl / .copy<input type="file" name="copybook" accept=".cpy,.cbl,.copy" required></label>
+        <label><strong>COBOL Source</strong>.cbl<input type="file" name="cobol_src" accept=".cbl" required></label>
+        <label><strong>Mapping</strong>.yaml / .yml<input type="file" name="mapping" accept=".yaml,.yml" required></label>
+        <label><strong>Runner</strong><select name="runner"><option value="native">Native Java</option><option value="spark">Spark Java</option></select></label>
+        <label class="full"><strong>Java Source</strong>directory with pom.xml<input type="file" name="java_src" webkitdirectory required></label>
+      </div>
+      <div class="actions">
+        <button type="submit" class="btn btn-primary">$ verify</button>
+        <button type="reset" class="btn btn-secondary">Clear</button>
+      </div>
+    </form>
+  </section>
+
+  <div id="status-area"></div>
+
+  <footer>
+    <span>verify-cli v0.2.0</span>
+    <span><a href="#">Docs</a> &middot; <a href="#">Architecture</a></span>
+  </footer>
+
 </div>
 <script src="/static/script.js"></script>
-</body></html>
+</body>
+</html>

web/worker.py

@@ -32,12 +32,20 @@ def main():
                 vr = run_pipeline(cfg, data["copybook"], data["cobol_src"],
                                   data["java_src"], data["mapping"])
 
+                fields = [{"name":fr.field_name,"status":fr.status,
+                           "cobol":fr.cobol_value,"java":fr.java_value,
+                           "suggestion":fr.suggestion} for fr in vr.field_results]
+
                 data["status"] = "done"
+                data["fields"] = fields
+                data["debug"] = vr.debug
                 data["result"] = {
                     "program": vr.program, "status": vr.status,
                     "matched": vr.fields_matched, "mismatched": vr.fields_mismatched,
                     "duration": vr.duration_s, "runner": vr.runner,
                 }
+                if vr.report_path and "BLOCKED" in vr.status:
+                    data["build_log"] = vr.report_path
                 tf.write_text(json.dumps(data))
 
             except Exception as e: