flonat-research 0.1.0

package/.scripts/generate-codex-agents-md.py

@@ -0,0 +1,217 @@
+#!/usr/bin/env python3
+"""Generate ~/.codex/AGENTS.md from Claude Code agents and rules.
+
+Reads agent definitions (.claude/agents/*.md) and rule files
+(.claude/rules/*.md), then produces a single AGENTS.md for Codex CLI.
+
+Usage:
+    uv run python .scripts/generate-codex-agents-md.py \
+        --agents-dir ~/.claude/agents \
+        --rules-dir ~/.claude/rules \
+        --output ~/.codex/AGENTS.md
+"""
+
+from __future__ import annotations
+
+import argparse
+import re
+from pathlib import Path
+
+
+def parse_frontmatter(text: str) -> tuple[dict[str, str], str]:
+    """Extract YAML frontmatter and body from markdown."""
+    if not text.startswith("---"):
+        return {}, text
+    end = text.find("---", 3)
+    if end == -1:
+        return {}, text
+    fm_text = text[3:end].strip()
+    body = text[end + 3 :].strip()
+    meta: dict[str, str] = {}
+    for line in fm_text.splitlines():
+        if ":" in line:
+            key, _, val = line.partition(":")
+            meta[key.strip()] = val.strip().strip('"').strip("'")
+    return meta, body
+
+
+def read_agent(path: Path) -> dict[str, str]:
+    """Read an agent file and return name + description + body."""
+    text = path.read_text(encoding="utf-8")
+    meta, body = parse_frontmatter(text)
+    return {
+        "name": meta.get("name", path.stem),
+        "description": meta.get("description", ""),
+        "body": body,
+        "filename": path.name,
+    }
+
+
+def read_rule(path: Path) -> dict[str, str]:
+    """Read a rule file and return its title + content."""
+    text = path.read_text(encoding="utf-8")
+    meta, body = parse_frontmatter(text)
+
+    # Extract first heading as title
+    title_match = re.search(r"^#\s+(?:Rule:\s*)?(.+)$", body, re.MULTILINE)
+    title = title_match.group(1).strip() if title_match else path.stem
+
+    return {
+        "title": title,
+        "body": body,
+        "filename": path.name,
+    }
+
+
+# Rules that are Claude-specific infrastructure and don't apply to Codex
+SKIP_RULES = {
+    "agents-vs-skills.md",  # Claude-specific agent/skill distinction
+    "break-the-glass.md",  # Claude infrastructure protection
+    "ignore-agents-md.md",  # Ironic — Codex *needs* AGENTS.md
+    "ignore-gemini-md.md",  # Claude-specific
+    "lean-claude-md.md",  # Claude-specific
+    "public-repo-sync.md",  # Claude infrastructure
+    "read-docs-first.md",  # Claude context system
+    "skill-outcome-logging.md",  # Claude skill logging
+    "severity-gradient.md",  # Claude review agents
+    "project-orchestration.md",  # Claude project commands
+}
+
+# Rules to include with full body (most important for behaviour)
+FULL_RULES = {
+    "scope-discipline.md",
+    "plan-first.md",
+    "data-sensitivity.md",
+    "overleaf-separation.md",
+    "design-before-results.md",
+    "no-hardcoded-results.md",
+    "latex-outdir.md",
+}
+
+# Agents to skip (Codex-specific ones don't make sense)
+SKIP_AGENTS = {
+    "codex-research.md",  # Meta — Codex reviewing itself
+}
+
+
+def generate_agents_md(
+    agents_dir: Path,
+    rules_dir: Path,
+) -> str:
+    """Generate the full AGENTS.md content."""
+    sections: list[str] = []
+
+    # ── Header ───────────────────────────────────────────────────────────
+    sections.append(
+        """\
+# Codex Instructions
+
+> Auto-generated by `sync-to-codex.sh` from Claude Code infrastructure.
+> Do not edit manually — changes will be overwritten on next sync.
+
+## Identity
+
+- **User:** the user
+- **Role:** PhD Researcher (Year 1)
+- **Affiliations:** [University 1], [University 2], [University 3] (research) + LSE (teaching)
+- **Research:** [your research areas]
+
+## Conventions
+
+- **LaTeX:** Build artifacts go to `out/` via `.latexmkrc`. Never leave `.aux`/`.log` in source.
+- **Python:** Always use `uv` — never bare `python` or `pip`.
+- **R:** Use `<-` for assignment, not `=`.
+- **Git:** Many repos are local-only (Dropbox-synced). Check `git remote -v` before pushing.
+- **Presentations:** Default to LaTeX Beamer.
+- **Data:** `data/raw/` is READ-ONLY. Never modify raw data files.
+"""
+    )
+
+    # ── Rules ────────────────────────────────────────────────────────────
+    rules_section = ["## Rules\n"]
+
+    rule_files = sorted(rules_dir.glob("*.md")) if rules_dir.is_dir() else []
+    included = 0
+    for rf in rule_files:
+        if rf.name in SKIP_RULES:
+            continue
+        rule = read_rule(rf)
+        if rf.name in FULL_RULES:
+            rules_section.append(f"### {rule['title']}\n")
+            # Strip the top-level heading from body (already used as ###)
+            body = re.sub(r"^#\s+.*$", "", rule["body"], count=1, flags=re.MULTILINE).strip()
+            # Downshift all headings by 2 levels (## → ####, ### → #####)
+            body = re.sub(r"^(#{2,})", lambda m: m.group(1) + "##", body, flags=re.MULTILINE)
+            rules_section.append(body)
+            rules_section.append("")
+        else:
+            # Summary only — first paragraph after the heading
+            body_lines = rule["body"].splitlines()
+            summary_lines: list[str] = []
+            past_heading = False
+            for line in body_lines:
+                if line.startswith("#"):
+                    if past_heading:
+                        break
+                    past_heading = True
+                    continue
+                if past_heading and line.strip():
+                    summary_lines.append(line.strip())
+                elif past_heading and not line.strip() and summary_lines:
+                    break
+            summary = " ".join(summary_lines) if summary_lines else rule["title"]
+            rules_section.append(f"- **{rule['title']}:** {summary}")
+        included += 1
+
+    sections.append("\n".join(rules_section))
+
+    # ── Agents (as reference) ────────────────────────────────────────────
+    agents_section = [
+        "\n## Available Review Agents\n",
+        "These agents are available in Claude Code for fresh-context review. ",
+        "When using Codex, their descriptions inform what each role does.\n",
+    ]
+
+    agent_files = sorted(agents_dir.glob("*.md")) if agents_dir.is_dir() else []
+    for af in agent_files:
+        if af.name in SKIP_AGENTS:
+            continue
+        agent = read_agent(af)
+        # Extract just the first line of description (before examples)
+        desc = agent["description"].split("\n")[0] if agent["description"] else ""
+        # Clean up escaped newlines from YAML
+        desc = desc.replace("\\n", " ").strip()
+        agents_section.append(f"- **{agent['name']}:** {desc}")
+
+    sections.append("\n".join(agents_section))
+
+    # ── Footer ───────────────────────────────────────────────────────────
+    sections.append(
+        """
+## Notion Databases
+
+| Database | ID |
+|----------|-----|
+| Tasks Tracker | `YOUR-TASKS-DATABASE-ID-HERE` |
+| Research Pipeline | `YOUR-PIPELINE-DATABASE-ID-HERE` |
+| Atlas (Topic Inventory) | `0a227f82-60f4-451a-a163-bff2ce8fa9c3` |
+"""
+    )
+
+    return "\n".join(sections)
+
+
+def main() -> None:
+    parser = argparse.ArgumentParser(description="Generate AGENTS.md for Codex CLI")
+    parser.add_argument("--agents-dir", type=Path, required=True)
+    parser.add_argument("--rules-dir", type=Path, required=True)
+    parser.add_argument("--output", type=Path, required=True)
+    args = parser.parse_args()
+
+    content = generate_agents_md(args.agents_dir, args.rules_dir)
+    args.output.write_text(content, encoding="utf-8")
+    print(f"Generated {args.output} ({len(content)} chars, {content.count(chr(10))} lines)")
+
+
+if __name__ == "__main__":
+    main()

package/.scripts/inbox

@@ -0,0 +1,194 @@
+#!/usr/bin/env python3
+"""
+Show tasks without due dates (inbox review).
+
+Usage:
+    inbox               # Show all inbox tasks
+    inbox --assign      # Interactively assign due dates
+"""
+
+import argparse
+import json
+import os
+import sys
+import urllib.request
+import urllib.error
+from datetime import datetime, timedelta
+
+
+NOTION_TOKEN = os.environ.get("NOTION_TOKEN")
+DATABASE_ID = "YOUR-TASKS-DATABASE-ID-HERE"
+
+
+def query_inbox():
+    """Query tasks without due dates."""
+    if not NOTION_TOKEN:
+        print("❌ NOTION_TOKEN not set")
+        sys.exit(1)
+
+    data = {
+        "filter": {
+            "and": [
+                {"property": "Status", "status": {"does_not_equal": "Done"}},
+                {"property": "Due date", "date": {"is_empty": True}}
+            ]
+        },
+        "sorts": [
+            {"property": "Priority", "direction": "ascending"},
+            {"timestamp": "created_time", "direction": "descending"}
+        ]
+    }
+
+    req = urllib.request.Request(
+        f"https://api.notion.com/v1/databases/{DATABASE_ID}/query",
+        data=json.dumps(data).encode("utf-8"),
+        headers={
+            "Authorization": f"Bearer {NOTION_TOKEN}",
+            "Content-Type": "application/json",
+            "Notion-Version": "2022-06-28",
+        },
+        method="POST",
+    )
+
+    try:
+        with urllib.request.urlopen(req) as response:
+            return json.loads(response.read().decode("utf-8"))
+    except urllib.error.HTTPError as e:
+        error_body = json.loads(e.read().decode("utf-8"))
+        print(f"❌ Error: {e.code}")
+        print(error_body.get("message", str(error_body)))
+        sys.exit(1)
+
+
+def get_property(page, prop_name, prop_type):
+    """Extract property value from a Notion page."""
+    prop = page.get("properties", {}).get(prop_name, {})
+
+    if prop_type == "title":
+        title_list = prop.get("title", [])
+        return title_list[0].get("plain_text", "") if title_list else ""
+    elif prop_type == "select":
+        select = prop.get("select")
+        return select.get("name", "") if select else ""
+    elif prop_type == "status":
+        status = prop.get("status")
+        return status.get("name", "") if status else ""
+    return ""
+
+
+def set_due_date(page_id, date_str):
+    """Set due date for a task."""
+    data = {
+        "properties": {
+            "Due date": {"date": {"start": date_str}}
+        }
+    }
+
+    req = urllib.request.Request(
+        f"https://api.notion.com/v1/pages/{page_id}",
+        data=json.dumps(data).encode("utf-8"),
+        headers={
+            "Authorization": f"Bearer {NOTION_TOKEN}",
+            "Content-Type": "application/json",
+            "Notion-Version": "2022-06-28",
+        },
+        method="PATCH",
+    )
+
+    try:
+        with urllib.request.urlopen(req) as response:
+            return True
+    except urllib.error.HTTPError as e:
+        return False
+
+
+def parse_date(date_str):
+    """Parse date input."""
+    today = datetime.now()
+    date_lower = date_str.lower().strip()
+
+    if date_lower in ["t", "today"]:
+        return today.strftime("%Y-%m-%d")
+    elif date_lower in ["tm", "tomorrow"]:
+        return (today + timedelta(days=1)).strftime("%Y-%m-%d")
+    elif date_lower in ["w", "week"]:
+        return (today + timedelta(days=7)).strftime("%Y-%m-%d")
+    elif date_lower in ["m", "month"]:
+        return (today + timedelta(days=30)).strftime("%Y-%m-%d")
+    elif date_lower in ["mon", "monday"]:
+        days = (7 - today.weekday()) % 7 or 7
+        return (today + timedelta(days=days)).strftime("%Y-%m-%d")
+    elif date_lower in ["fri", "friday"]:
+        days = (4 - today.weekday()) % 7 or 7
+        return (today + timedelta(days=days)).strftime("%Y-%m-%d")
+    else:
+        return date_str
+
+
+def format_task(page, index=None):
+    """Format a task for display."""
+    title = get_property(page, "Task name", "title")
+    priority = get_property(page, "Priority", "select")
+    project = get_property(page, "Project", "select")
+
+    priority_emoji = {"High": "🔴", "Medium": "🟡", "Low": "🟢"}.get(priority, "⚪")
+
+    if index is not None:
+        line = f"{index:2}. {priority_emoji} {title}"
+    else:
+        line = f"{priority_emoji} {title}"
+
+    if project:
+        line += f" [{project}]"
+
+    return line
+
+
+def main():
+    parser = argparse.ArgumentParser(description="Inbox review - tasks without due dates")
+    parser.add_argument("--assign", action="store_true", help="Interactively assign due dates")
+
+    args = parser.parse_args()
+
+    result = query_inbox()
+    pages = result.get("results", [])
+
+    if not pages:
+        print("📥 Inbox empty! All tasks have due dates.")
+        return
+
+    print(f"📥 Inbox ({len(pages)} tasks without due dates)")
+    print("-" * 50)
+
+    if args.assign:
+        print("Shortcuts: t=today, tm=tomorrow, w=week, m=month, fri=friday, s=skip, q=quit\n")
+
+        for i, page in enumerate(pages, 1):
+            print(format_task(page, i))
+            try:
+                date_input = input("   Due date: ").strip()
+
+                if date_input.lower() == 'q':
+                    print("Done.")
+                    break
+                elif date_input.lower() == 's' or not date_input:
+                    continue
+                else:
+                    date_str = parse_date(date_input)
+                    if set_due_date(page["id"], date_str):
+                        print(f"   ✅ Set to {date_str}\n")
+                    else:
+                        print(f"   ❌ Failed to set date\n")
+            except KeyboardInterrupt:
+                print("\nDone.")
+                break
+    else:
+        for page in pages:
+            print(format_task(page))
+
+        print()
+        print("Run 'inbox --assign' to set due dates interactively.")
+
+
+if __name__ == "__main__":
+    main()