superlab 0.1.65 → 0.1.67

package/package-assets/shared/lab/.managed/scripts/validate_manuscript_delivery.py

@@ -39,6 +39,29 @@ REQUIRED_TABLE_NOTE_MARKERS = (
 )
 WIDTH_CONTROL_NOTE_MARKER = "% Width control:"
 TABLE_ABBREVIATION_EXCEPTIONS = {"TODO", "TBD"}
+SERVICE_STYLE_PHRASES = (
+    "user asked",
+    "the user asked",
+    "as requested by the user",
+    "let me explain",
+    "i will explain",
+    "用户说",
+    "用户要求",
+    "按你的要求",
+    "我来解释",
+    "我会说明",
+)
+WORKFLOW_ONLY_MANUSCRIPT_PHRASES = (
+    "Figure intent:",
+    "Asset intent:",
+    "figure intent:",
+    "asset intent:",
+    "图的意图",
+    "资产意图",
+    "占位符",
+    "workflow-language",
+    "translation layer",
+)
 
 
 def parse_args():
@@ -53,6 +76,15 @@ def read_text(path: Path) -> str:
     return path.read_text(encoding="utf-8")
 
 
+def strip_latex_comments(text: str) -> str:
+    return re.sub(r"%.*", "", text)
+
+
+def contains_any(text: str, needles: tuple[str, ...]) -> bool:
+    lowered = text.lower()
+    return any(needle.lower() in lowered for needle in needles)
+
+
 def find_workflow_config(start_path: Path) -> Path | None:
     search_roots = [start_path, *start_path.parents]
     for root in search_roots:
@@ -149,54 +181,6 @@ def extract_table_note(text: str) -> str:
     return "\n".join(line.strip() for line in text.splitlines() if line.strip().startswith("%"))
 
 
-def extract_tabular_specs(text: str) -> list[str]:
-    return re.findall(r"\\begin\{tabular\}(?:\[[^\]]*\])?\{([^}]*)\}", text)
-
-
-def extract_tabular_bodies(text: str) -> list[str]:
-    return [
-        match.group(1)
-        for match in re.finditer(
-            r"\\begin\{tabular\}(?:\[[^\]]*\])?\{[^}]*\}([\s\S]*?)\\end\{tabular\}",
-            text,
-        )
-    ]
-
-
-def strip_latex_for_table_cell(cell: str) -> str:
-    cell = re.sub(r"%.*", "", cell)
-    cell = re.sub(r"\\(?:textbf|mathbf|emph)\{([^}]*)\}", r"\1", cell)
-    cell = re.sub(r"\\[A-Za-z@*]+(?:\[[^\]]*\])?", " ", cell)
-    cell = cell.replace("$", " ")
-    return cell.strip()
-
-
-def check_numeric_precision_consistency(text: str, issues: list[str], label: str):
-    for body in extract_tabular_bodies(text):
-        column_precisions: dict[int, set[int]] = {}
-        rows = re.split(r"\\\\", body)
-        for row in rows:
-            row = re.sub(r"\\(?:toprule|midrule|bottomrule|cmidrule)(?:\{[^}]*\})?", " ", row)
-            if "&" not in row:
-                continue
-            cells = [strip_latex_for_table_cell(cell) for cell in row.split("&")]
-            for index, cell in enumerate(cells):
-                matches = re.findall(r"(?<![A-Za-z])[-+]?\d+\.(\d+)", cell)
-                if not matches:
-                    continue
-                column_precisions.setdefault(index, set()).update(len(match) for match in matches)
-        inconsistent_columns = [
-            str(index + 1)
-            for index, precisions in sorted(column_precisions.items())
-            if len(precisions) > 1
-        ]
-        if inconsistent_columns:
-            issues.append(
-                f"{label} has inconsistent numeric precision in column(s): {', '.join(inconsistent_columns)}"
-            )
-            return
-
-
 def detect_uppercase_abbreviations(text: str) -> set[str]:
     return {
         token
@@ -258,16 +242,8 @@ def check_table_file(path: Path, issues: list[str], label: str):
         issues.append(f"{label} must contain a table environment")
     if r"\caption{" not in text or r"\label{" not in text:
         issues.append(f"{label} must contain both caption and label")
-    first_caption = text.find(r"\caption{")
-    first_tabular = text.find(r"\begin{tabular}")
-    if first_caption != -1 and first_tabular != -1 and first_caption > first_tabular:
-        issues.append(f"{label} should place its caption before the tabular body")
     if not all(token in text for token in (r"\toprule", r"\midrule", r"\bottomrule")):
         issues.append(f"{label} must use booktabs structure")
-    if any("|" in spec for spec in extract_tabular_specs(text)) or r"\vline" in text:
-        issues.append(f"{label} must not use vertical table rules; use booktabs spacing instead")
-    if r"\hline" in text or r"\cline" in text:
-        issues.append(f"{label} must not mix legacy \\hline/\\cline rules with booktabs tables")
     if not all(marker in text for marker in REQUIRED_TABLE_NOTE_MARKERS):
         issues.append(f"{label} must include a local table note scaffold")
     caption_text = extract_caption(text)
@@ -290,7 +266,6 @@ def check_table_file(path: Path, issues: list[str], label: str):
             continue
         if value < 3.0:
             issues.append(f"{label} sets \\tabcolsep below the safe range for paper-facing main tables")
-    check_numeric_precision_consistency(text, issues, label)
 
 
 def check_figure_file(path: Path, issues: list[str], label: str):
@@ -309,8 +284,11 @@ def check_figure_file(path: Path, issues: list[str], label: str):
         issues.append(f"{label} must contain a figure environment")
     if r"\caption{" not in text or r"\label{" not in text:
         issues.append(f"{label} must contain both caption and label")
-    if "Figure intent:" not in text and "图意图：" not in text:
-        issues.append(f"{label} must explain figure intent")
+    paper_facing_text = strip_latex_comments(text)
+    if contains_any(paper_facing_text, SERVICE_STYLE_PHRASES):
+        issues.append(f"{label} contains service-style or AI-assistant meta language")
+    if contains_any(paper_facing_text, WORKFLOW_ONLY_MANUSCRIPT_PHRASES):
+        issues.append(f"{label} contains workflow-only placeholder language in paper-facing text")
 
 
 def check_analysis_asset(path: Path, issues: list[str]):
@@ -322,8 +300,11 @@ def check_analysis_asset(path: Path, issues: list[str]):
         issues.append("analysis/analysis-asset.tex must contain a table or figure environment")
     if r"\caption{" not in text or r"\label{" not in text:
         issues.append("analysis/analysis-asset.tex must contain both caption and label")
-    if "Asset intent:" not in text and "资产意图：" not in text:
-        issues.append("analysis/analysis-asset.tex must explain asset intent")
+    paper_facing_text = strip_latex_comments(text)
+    if contains_any(paper_facing_text, SERVICE_STYLE_PHRASES):
+        issues.append("analysis/analysis-asset.tex contains service-style or AI-assistant meta language")
+    if contains_any(paper_facing_text, WORKFLOW_ONLY_MANUSCRIPT_PHRASES):
+        issues.append("analysis/analysis-asset.tex contains workflow-only placeholder language in paper-facing text")
 
 
 def require_section_reference(section_text: str, label: str | None, issues: list[str], section_name: str, asset_name: str):

package/package-assets/shared/lab/.managed/scripts/validate_paper_plan.py

@@ -82,6 +82,7 @@ TABLE_DETAIL_FIELDS = (
 FIGURE_DETAIL_FIELDS = (
     "Asset file",
     "Section",
+    "Figure role",
     "Figure intent",
     "Evidence",
     "Status",
@@ -95,6 +96,7 @@ ANALYSIS_DETAIL_FIELDS = (
     "Asset file",
     "Asset type",
     "Section",
+    "Asset role",
     "Asset intent",
     "Evidence",
     "Status",

package/package-assets/shared/lab/.managed/scripts/validate_reference_consumption.py

@@ -0,0 +1,230 @@
+#!/usr/bin/env python3
+from __future__ import annotations
+
+import argparse
+import re
+import sys
+from pathlib import Path
+
+from paper_topology import find_project_root
+
+
+SECTION_CHOICES = (
+    "abstract",
+    "introduction",
+    "related-work",
+    "method",
+    "experiments",
+    "conclusion",
+)
+
+REQUIRED_HEADINGS = (
+    "## Sources",
+    "## Adopted Structure Slots",
+    "## Rejected or Waived Slots",
+    "## Section Mapping",
+    "## Asset Mapping",
+    "## Reuse Boundary",
+)
+
+EXPERIMENT_CORE_SLOTS = {
+    "dataset_description": ("dataset_description", "dataset statistics", "dataset", "datasets", "数据集"),
+    "split_protocol": ("split_protocol", "split protocol", "split", "train/test", "划分", "切分"),
+    "baseline_setup": ("baseline_setup", "baseline", "baselines", "comparator", "comparators", "基线", "对比方法"),
+    "metric_definition": ("metric_definition", "metric", "metrics", "ranking", "指标", "排序"),
+    "main_results": ("main_results", "main result", "main results", "primary comparison", "主结果", "主要结果"),
+}
+
+EXPERIMENT_WAIVABLE_SLOTS = {
+    "ablation": ("ablation", "component analysis", "消融", "组件分析"),
+    "sensitivity": ("sensitivity", "robustness", "trade-off", "敏感性", "稳健性"),
+    "implementation_details": ("implementation_details", "implementation", "reproducibility", "实现", "复现"),
+}
+
+SERVICE_STYLE_PHRASES = (
+    "user asked",
+    "the user asked",
+    "as requested by the user",
+    "i will explain",
+    "let me explain",
+    "below i",
+    "用户说",
+    "用户要求",
+    "按你的要求",
+    "我来解释",
+    "我会说明",
+    "下面我",
+    "这版",
+    "已完成",
+    "已按",
+)
+
+WORKFLOW_ONLY_PHRASES = (
+    "figure intent",
+    "asset intent",
+    "placeholder",
+    "workflow-language",
+    "review layer",
+    "translation layer",
+    "图的意图",
+    "资产意图",
+    "占位符",
+    "工作流语言",
+    "同步到",
+)
+
+
+def parse_args() -> argparse.Namespace:
+    parser = argparse.ArgumentParser(
+        description="Validate that reference-paper structure was consumed as structure, not as shallow prose polish."
+    )
+    parser.add_argument("--section", required=True, choices=SECTION_CHOICES)
+    parser.add_argument("--section-file", required=True)
+    parser.add_argument("--consumption-plan", help="Path to .lab/writing/reference-patterns/consumption-plan/<section>.md")
+    parser.add_argument("--mode", required=True, choices=("draft", "final"))
+    return parser.parse_args()
+
+
+def read_text(path: Path) -> str:
+    return path.read_text(encoding="utf-8")
+
+
+def contains_any(text: str, phrases: tuple[str, ...]) -> bool:
+    lowered = text.lower()
+    return any(phrase.lower() in lowered for phrase in phrases)
+
+
+def strip_latex_comments(text: str) -> str:
+    return re.sub(r"%.*", "", text)
+
+
+def infer_consumption_plan(section_file: Path, section: str) -> Path | None:
+    project_root = find_project_root(section_file.resolve())
+    if project_root is None:
+        return None
+    return project_root / ".lab" / "writing" / "reference-patterns" / "consumption-plan" / f"{section}.md"
+
+
+def section_block(text: str, heading: str) -> str:
+    match = re.search(
+        rf"^##\s+{re.escape(heading.removeprefix('## ').strip())}\s*$([\s\S]*?)(?=^##\s+|\Z)",
+        text,
+        flags=re.MULTILINE,
+    )
+    return match.group(1) if match else ""
+
+
+def nonempty_bullets(block: str) -> list[str]:
+    return [
+        line.strip()
+        for line in block.splitlines()
+        if line.strip().startswith("-") and line.strip().lower() not in {"- none", "- n/a", "- na"}
+    ]
+
+
+def line_mentions_any(line: str, aliases: tuple[str, ...]) -> bool:
+    lowered = line.lower()
+    return any(alias.lower() in lowered for alias in aliases)
+
+
+def slot_is_mapped(plan_text: str, aliases: tuple[str, ...]) -> bool:
+    mapping_blocks = "\n".join(
+        [
+            section_block(plan_text, "Adopted Structure Slots"),
+            section_block(plan_text, "Section Mapping"),
+            section_block(plan_text, "Asset Mapping"),
+        ]
+    )
+    for line in mapping_blocks.splitlines():
+        if line_mentions_any(line, aliases) and ("->" in line or ":" in line):
+            return True
+    return False
+
+
+def slot_is_waived(plan_text: str, aliases: tuple[str, ...]) -> bool:
+    waiver_block = section_block(plan_text, "Rejected or Waived Slots")
+    for line in waiver_block.splitlines():
+        lowered = line.lower()
+        if line_mentions_any(line, aliases) and any(token in lowered for token in ("waiv", "not used", "not applicable", "不采用", "不适用", "跳过")):
+            return True
+    return False
+
+
+def validate_plan_shape(plan_text: str, issues: list[str]) -> None:
+    for heading in REQUIRED_HEADINGS:
+        if heading not in plan_text:
+            issues.append(f"reference consumption plan is missing required heading: {heading}")
+
+    source_block = section_block(plan_text, "Sources")
+    if len(nonempty_bullets(source_block)) == 0:
+        issues.append("reference consumption plan must list at least one reference source")
+
+    boundary = section_block(plan_text, "Reuse Boundary").lower()
+    required_boundary_terms = ("structure", "wording", "claims", "metrics", "captions", "conclusions")
+    if not all(term in boundary for term in required_boundary_terms):
+        issues.append(
+            "reuse boundary must explicitly say to reuse structure only and not copy wording, claims, metrics, captions, or conclusions"
+        )
+
+
+def validate_experiment_slots(plan_text: str, issues: list[str]) -> None:
+    for slot, aliases in EXPERIMENT_CORE_SLOTS.items():
+        if not slot_is_mapped(plan_text, aliases):
+            issues.append(f"experiments reference consumption must map required slot: {slot}")
+
+    for slot, aliases in EXPERIMENT_WAIVABLE_SLOTS.items():
+        if not slot_is_mapped(plan_text, aliases) and not slot_is_waived(plan_text, aliases):
+            issues.append(f"experiments reference consumption must map or explicitly waive slot: {slot}")
+
+
+def validate_section_text(section_text: str, issues: list[str]) -> None:
+    prose_text = strip_latex_comments(section_text)
+    if contains_any(prose_text, SERVICE_STYLE_PHRASES):
+        issues.append("section prose contains service-style or AI-assistant meta language; rewrite as paper-facing academic prose")
+    if contains_any(prose_text, WORKFLOW_ONLY_PHRASES):
+        issues.append("section prose contains workflow-only placeholder language; move authoring notes out of the manuscript")
+
+
+def main() -> int:
+    args = parse_args()
+    section_path = Path(args.section_file)
+    if not section_path.exists():
+        print(f"section file does not exist: {section_path}", file=sys.stderr)
+        return 1
+
+    plan_path = Path(args.consumption_plan) if args.consumption_plan else infer_consumption_plan(section_path, args.section)
+    if plan_path is None or not plan_path.exists():
+        message = (
+            "missing reference consumption plan; create "
+            f".lab/writing/reference-patterns/consumption-plan/{args.section}.md before reference-guided deep writing"
+        )
+        if args.mode == "draft":
+            print(f"WARNING: {message}")
+            return 0
+        print(message, file=sys.stderr)
+        return 1
+
+    issues: list[str] = []
+    plan_text = read_text(plan_path)
+    section_text = read_text(section_path)
+    validate_plan_shape(plan_text, issues)
+    if args.section == "experiments":
+        validate_experiment_slots(plan_text, issues)
+    validate_section_text(section_text, issues)
+
+    if not issues:
+        print("reference consumption is valid")
+        return 0
+
+    if args.mode == "draft":
+        for issue in issues:
+            print(f"WARNING: {issue}")
+        return 0
+
+    for issue in issues:
+        print(issue, file=sys.stderr)
+    return 1
+
+
+if __name__ == "__main__":
+    raise SystemExit(main())

package/package-assets/shared/lab/.managed/scripts/validate_section_draft.py

@@ -198,9 +198,49 @@ SECTION_STYLE_WARNINGS = {
     ],
 }
 
+SERVICE_STYLE_PHRASES = (
+    "user asked",
+    "the user asked",
+    "as requested by the user",
+    "let me explain",
+    "i will explain",
+    "below i",
+    "用户说",
+    "用户要求",
+    "按你的要求",
+    "我来解释",
+    "我会说明",
+    "下面我",
+    "这版",
+    "已完成",
+    "已按",
+)
+
+WORKFLOW_ONLY_MANUSCRIPT_PHRASES = (
+    "figure intent",
+    "asset intent",
+    "placeholder",
+    "workflow-language",
+    "translation layer",
+    "review layer",
+    "图的意图",
+    "资产意图",
+    "占位符",
+    "工作流语言",
+    "同步到",
+)
+
 
 def check_common_section_gate_risks(text: str, issues: list[str]):
     prose_text = strip_latex_commands(text)
+    if contains_any(prose_text, SERVICE_STYLE_PHRASES):
+        issues.append(
+            "service-style or AI-assistant meta language appears in reader-facing prose; rewrite it as academic manuscript text"
+        )
+    if contains_any(prose_text, WORKFLOW_ONLY_MANUSCRIPT_PHRASES):
+        issues.append(
+            "workflow-only placeholder language appears in reader-facing prose; move authoring notes out of the manuscript"
+        )
     if re.search(r"\b[a-z0-9]+(?:_[a-z0-9]+)+\b", prose_text):
         issues.append(
             "reader-facing prose appears to contain internal identifier-like tokens; map them once for the reader and move them back out of prose before more polishing"
@@ -302,6 +342,16 @@ def check_neighbor_asset_files(section: str, section_path: Path, issues: list[st
             issues.append(
                 f"{section} section is missing the required paper-layer asset file: {asset_path.as_posix()}"
             )
+            continue
+        asset_text = strip_latex_commands(read_text(asset_path))
+        if contains_any(asset_text, SERVICE_STYLE_PHRASES):
+            issues.append(
+                f"{asset_path.as_posix()} contains service-style or AI-assistant meta language; rewrite it as paper-facing asset text"
+            )
+        if contains_any(asset_text, WORKFLOW_ONLY_MANUSCRIPT_PHRASES):
+            issues.append(
+                f"{asset_path.as_posix()} contains workflow-only placeholder language; move authoring notes out of captions and paper-facing asset text"
+            )
 
 
 def check_paper_topology_targeting(section_path: Path, issues: list[str]):
@@ -504,100 +554,6 @@ def check_method(text: str, issues: list[str]):
         issues.append("method should explain the technical advantage")
 
 
-def has_performance_claim(text: str) -> bool:
-    return contains_any(
-        text,
-        (
-            "outperform",
-            "outperforms",
-            "improve",
-            "improves",
-            "improved",
-            "gain",
-            "gains",
-            "better",
-            "stronger",
-            "superior",
-            "state-of-the-art",
-            "sota",
-            "reduce",
-            "reduces",
-            "降低",
-            "提升",
-            "优于",
-            "超过",
-            "更好",
-            "增益",
-        ),
-    )
-
-
-def has_numeric_or_table_evidence(text: str) -> bool:
-    if re.search(r"\b\d+\.\d+\b", text):
-        return True
-    if re.search(r"\b\d+(?:\.\d+)?\s*(?:%|pp|points?|AUUC|Qini|AUC|F1)\b", text, flags=re.IGNORECASE):
-        return True
-    if r"\pm" in text:
-        return True
-    return bool(
-        re.search(r"\\(?:auto|c|C)?ref\{(?:tab|fig):", text)
-        or re.search(r"\b(?:Table|Figure|Fig\.|表|图)~?\\ref\{", text)
-    )
-
-
-def has_generic_comparator_without_anchor(text: str) -> bool:
-    generic_comparator = contains_any(
-        text,
-        (
-            "previous methods",
-            "prior methods",
-            "existing methods",
-            "several baselines",
-            "the baselines",
-            "baseline suite",
-            "previous work",
-            "prior work",
-            "现有方法",
-            "已有方法",
-            "若干基线",
-            "基线集合",
-        ),
-    )
-    if not generic_comparator:
-        return False
-    if r"\cite{" in text or r"\citet{" in text or r"\citep{" in text:
-        return False
-    return not bool(re.search(r"\b[A-Z][A-Za-z0-9-]{2,}(?:\s*,\s*[A-Z][A-Za-z0-9-]{2,})+", text))
-
-
-def has_repeated_split_protocol(text: str) -> bool:
-    return bool(
-        re.search(r"\b\d+\s+(?:random\s+)?(?:splits|seeds|runs)\b", text, flags=re.IGNORECASE)
-        or re.search(r"\bacross\s+(?:random\s+)?(?:splits|seeds|runs)\b", text, flags=re.IGNORECASE)
-        or re.search(r"\b重复\s*\d+\s*次", text)
-    )
-
-
-def has_variance_report(text: str) -> bool:
-    return contains_any(
-        text,
-        (
-            r"\pm",
-            "standard deviation",
-            "std",
-            "confidence interval",
-            "confidence intervals",
-            "ci",
-            "variance",
-            "mean",
-            "平均",
-            "标准差",
-            "置信区间",
-            "方差",
-        ),
-    )
-
-
 def check_experiments(text: str, issues: list[str]):
     if not contains_any(
         text,
@@ -625,18 +581,6 @@ def check_experiments(text: str, issues: list[str]):
         ),
     ):
         issues.append("experiments should include benchmark scene notes")
-    if has_performance_claim(text) and not has_numeric_or_table_evidence(text):
-        issues.append(
-            "experiment performance claims should tie to concrete metric or numeric evidence instead of prose-only claims"
-        )
-    if has_generic_comparator_without_anchor(text):
-        issues.append(
-            "experiments use generic comparator names; name the comparator family, table anchor, or citations before more polish"
-        )
-    if has_repeated_split_protocol(text) and not has_variance_report(text):
-        issues.append(
-            "repeated split or seed protocol should report variance, confidence intervals, or an explicit variance disposition"
-        )
 
 
 def check_conclusion(text: str, issues: list[str]):

package/package-assets/shared/lab/.managed/templates/paper-figure.tex

@@ -1,6 +1,7 @@
 \begin{figure}[t]
 \centering
 \fbox{\rule{0pt}{1.2in}\rule{0.9\linewidth}{0pt}}
-\caption{Figure title. Figure intent: explain what this figure should show and why the reader needs it.}
+% Authoring note: record what this figure should show and why the reader needs it in the paper plan or write-iteration artifact, not in the caption.
+\caption{Figure title. Replace with a paper-facing caption that states the visual content and the supported claim.}
 \label{fig:placeholder}
 \end{figure}

package/package-assets/shared/lab/.managed/templates/paper-plan.md

@@ -51,19 +51,19 @@
 - Problem setting or teaser figure:
   - Asset file:
   - Section:
-  - Figure intent:
+  - Figure role:
   - Evidence:
   - Status:
 - Method overview figure:
   - Asset file:
   - Section:
-  - Figure intent:
+  - Figure role:
   - Evidence:
   - Status:
 - Results overview figure:
   - Asset file:
   - Section:
-  - Figure intent:
+  - Figure role:
   - Evidence:
   - Status:
 
@@ -73,7 +73,7 @@
   - Asset file:
   - Asset type:
   - Section:
-  - Asset intent:
+  - Asset role:
   - Evidence:
   - Status:
 

package/package-assets/shared/lab/.managed/templates/reference-consumption-plan.md

@@ -0,0 +1,34 @@
+# Reference Consumption Plan
+
+## Sources
+
+- Source:
+
+## Adopted Structure Slots
+
+- slot_name -> target subsection / paragraph / asset
+
+## Rejected or Waived Slots
+
+- slot_name -> waiver reason
+
+## Section Mapping
+
+- slot_name -> current-paper section or subsection
+
+## Paragraph Role Mapping
+
+- paragraph_role -> current-paper paragraph
+
+## Asset Mapping
+
+- asset_role -> current-paper figure/table/analysis asset
+
+## Reuse Boundary
+
+- Reuse structure, section order, paragraph roles, asset function, placement logic, and bridge logic only.
+- Do not copy wording, claims, metrics, captions, or conclusions from reference papers.
+
+## Validation
+
+- `validate_reference_consumption.py` result: