worklog-for-claude-code 0.1.0__tar.gz → 0.2.0__tar.gz

{worklog_for_claude_code-0.1.0/worklog_for_claude_code.egg-info → worklog_for_claude_code-0.2.0}/PKG-INFO

@@ -1,14 +1,14 @@
 Metadata-Version: 2.4
 Name: worklog-for-claude-code
-Version: 0.1.0
-Summary: Daily work log and token report across Claude Code accounts and devices, synced through your own private git repo
+Version: 0.2.0
+Summary: Daily work log and token report across Claude Code accounts and devices, plus imported claude.ai chats, synced through your own private git repo
 Author: Urvin
 License-Expression: MIT
 Project-URL: Homepage, https://github.com/us-urvin/worklog-for-claude-code
 Project-URL: Source, https://github.com/us-urvin/worklog-for-claude-code
 Project-URL: Issues, https://github.com/us-urvin/worklog-for-claude-code/issues
 Project-URL: Changelog, https://github.com/us-urvin/worklog-for-claude-code/blob/main/CHANGELOG.md
-Keywords: claude-code,claude,worklog,time-tracking,token-usage,developer-tools,daily-report
+Keywords: claude-code,claude,claude-ai,worklog,time-tracking,token-usage,developer-tools,daily-report
 Classifier: Development Status :: 4 - Beta
 Classifier: Environment :: Console
 Classifier: Environment :: Web Environment
@@ -40,7 +40,7 @@ Dynamic: license-file
 
 [![CI](https://github.com/us-urvin/worklog-for-claude-code/actions/workflows/ci.yml/badge.svg)](https://github.com/us-urvin/worklog-for-claude-code/actions/workflows/ci.yml)
 [![PyPI](https://img.shields.io/pypi/v/worklog-for-claude-code)](https://pypi.org/project/worklog-for-claude-code/)
-[![License: MIT](https://img.shields.io/badge/license-MIT-blue)](LICENSE)
+[![License: MIT](https://img.shields.io/badge/license-MIT-blue)](https://github.com/us-urvin/worklog-for-claude-code/blob/main/LICENSE)
 
 > Independent community project, not affiliated with or endorsed by Anthropic.
 
@@ -53,6 +53,7 @@ You use Claude Code all day. At the end of it, this tool gives you a simple page
 - ✅ **What you worked on**: a short summary for each project
 - 📊 **How many tokens you used**: per day, per account and per model
 - 💻 **All in one place**: even if you use several computers or several Claude accounts
+- 💬 **Your claude.ai chats too** (optional): import claude.ai's data export to see chats next to your coding work
 
 It works in the background. You don't have to do anything after setup.
 
@@ -83,30 +84,31 @@ claude-worklog install-service
 
 Done. Your work log fills up as you use Claude Code.
 
-> Using more than one computer? Put the data folder in a **private** GitHub/GitLab repo so all your machines share it. See [Full setup](docs/GUIDE.md#full-setup).
+> Using more than one computer? Put the data folder in a **private** GitHub/GitLab repo so all your machines share it. See [Full setup](https://github.com/us-urvin/worklog-for-claude-code/blob/main/docs/GUIDE.md#full-setup).
 
 ## Is my data private?
 
-Yes.
+Mostly yes, with a few things you should know.
 
 - 🔒 **Everything stays on your computer**, plus your own private git repo if you set one up. No cloud service, no tracking.
-- ✂️ **Only short notes are saved**: your prompts (shortened), token counts, file names and commit messages. Never your code, and never Claude's replies.
-- 🧹 **Secrets are removed**: API keys, passwords, tokens and e-mail addresses are cleaned out before saving.
+- ✂️ **Only short notes are saved**: the first 300 characters of each prompt, token counts, file names and commit messages. Your files and Claude's replies are never saved, but a code snippet you paste into a prompt can end up in those 300 characters.
+- 🧹 **Common secrets are hidden (best effort)**: well-known API key formats, `password=…` / `token=…` pairs, e-mail addresses and IP addresses are replaced with `[REDACTED]` before saving. This can miss things, such as a password written in a normal sentence, so keep your data repo **private**.
 - 🤖 **Optional AI summaries** use your own Claude account (`claude -p`). Turn them off in Settings to keep everything fully local.
+- 💬 **Imported claude.ai chats** keep only chat titles and short redacted snippets of your own messages. Claude's replies, chat summaries and attachments are never saved.
 
-[More about privacy →](docs/GUIDE.md#privacy)
+[More about privacy →](https://github.com/us-urvin/worklog-for-claude-code/blob/main/docs/GUIDE.md#privacy)
 
 ## Good to know
 
-- Works with **Claude Code only**, not claude.ai chat.
+- **Claude Code is tracked automatically.** claude.ai chats have no live connection: you import its data export on the **Sources** page (upload it, or let worklog pick it up from your Downloads folder). Exports have no token counts. See [claude.ai chats](https://github.com/us-urvin/worklog-for-claude-code/blob/main/docs/GUIDE.md#claudeai-chats).
 - **Linux only** for now (it uses systemd).
 - It reads Claude Code's local log files. Their format isn't official, so a Claude Code update could break it for a while.
 
 ## More
 
-- 📖 [Full guide](docs/GUIDE.md): several computers, several accounts, all commands, uninstalling
-- 🐛 [Report a bug](https://github.com/us-urvin/worklog-for-claude-code/issues/new/choose) · 🔐 [Security](SECURITY.md) · 🤝 [Contributing](CONTRIBUTING.md) · 📝 [Changelog](CHANGELOG.md)
+- 📖 [Full guide](https://github.com/us-urvin/worklog-for-claude-code/blob/main/docs/GUIDE.md): several computers, several accounts, all commands, uninstalling
+- 🐛 [Report a bug](https://github.com/us-urvin/worklog-for-claude-code/issues/new/choose) · 🔐 [Security](https://github.com/us-urvin/worklog-for-claude-code/blob/main/SECURITY.md) · 🤝 [Contributing](https://github.com/us-urvin/worklog-for-claude-code/blob/main/CONTRIBUTING.md) · 📝 [Changelog](https://github.com/us-urvin/worklog-for-claude-code/blob/main/CHANGELOG.md)
 
 ## License
 
-[MIT](LICENSE) © 2026 Urvin
+[MIT](https://github.com/us-urvin/worklog-for-claude-code/blob/main/LICENSE) © 2026 Urvin

{worklog_for_claude_code-0.1.0 → worklog_for_claude_code-0.2.0}/README.md

@@ -4,7 +4,7 @@
 
 [![CI](https://github.com/us-urvin/worklog-for-claude-code/actions/workflows/ci.yml/badge.svg)](https://github.com/us-urvin/worklog-for-claude-code/actions/workflows/ci.yml)
 [![PyPI](https://img.shields.io/pypi/v/worklog-for-claude-code)](https://pypi.org/project/worklog-for-claude-code/)
-[![License: MIT](https://img.shields.io/badge/license-MIT-blue)](LICENSE)
+[![License: MIT](https://img.shields.io/badge/license-MIT-blue)](https://github.com/us-urvin/worklog-for-claude-code/blob/main/LICENSE)
 
 > Independent community project, not affiliated with or endorsed by Anthropic.
 
@@ -17,6 +17,7 @@ You use Claude Code all day. At the end of it, this tool gives you a simple page
 - ✅ **What you worked on**: a short summary for each project
 - 📊 **How many tokens you used**: per day, per account and per model
 - 💻 **All in one place**: even if you use several computers or several Claude accounts
+- 💬 **Your claude.ai chats too** (optional): import claude.ai's data export to see chats next to your coding work
 
 It works in the background. You don't have to do anything after setup.
 
@@ -47,30 +48,31 @@ claude-worklog install-service
 
 Done. Your work log fills up as you use Claude Code.
 
-> Using more than one computer? Put the data folder in a **private** GitHub/GitLab repo so all your machines share it. See [Full setup](docs/GUIDE.md#full-setup).
+> Using more than one computer? Put the data folder in a **private** GitHub/GitLab repo so all your machines share it. See [Full setup](https://github.com/us-urvin/worklog-for-claude-code/blob/main/docs/GUIDE.md#full-setup).
 
 ## Is my data private?
 
-Yes.
+Mostly yes, with a few things you should know.
 
 - 🔒 **Everything stays on your computer**, plus your own private git repo if you set one up. No cloud service, no tracking.
-- ✂️ **Only short notes are saved**: your prompts (shortened), token counts, file names and commit messages. Never your code, and never Claude's replies.
-- 🧹 **Secrets are removed**: API keys, passwords, tokens and e-mail addresses are cleaned out before saving.
+- ✂️ **Only short notes are saved**: the first 300 characters of each prompt, token counts, file names and commit messages. Your files and Claude's replies are never saved, but a code snippet you paste into a prompt can end up in those 300 characters.
+- 🧹 **Common secrets are hidden (best effort)**: well-known API key formats, `password=…` / `token=…` pairs, e-mail addresses and IP addresses are replaced with `[REDACTED]` before saving. This can miss things, such as a password written in a normal sentence, so keep your data repo **private**.
 - 🤖 **Optional AI summaries** use your own Claude account (`claude -p`). Turn them off in Settings to keep everything fully local.
+- 💬 **Imported claude.ai chats** keep only chat titles and short redacted snippets of your own messages. Claude's replies, chat summaries and attachments are never saved.
 
-[More about privacy →](docs/GUIDE.md#privacy)
+[More about privacy →](https://github.com/us-urvin/worklog-for-claude-code/blob/main/docs/GUIDE.md#privacy)
 
 ## Good to know
 
-- Works with **Claude Code only**, not claude.ai chat.
+- **Claude Code is tracked automatically.** claude.ai chats have no live connection: you import its data export on the **Sources** page (upload it, or let worklog pick it up from your Downloads folder). Exports have no token counts. See [claude.ai chats](https://github.com/us-urvin/worklog-for-claude-code/blob/main/docs/GUIDE.md#claudeai-chats).
 - **Linux only** for now (it uses systemd).
 - It reads Claude Code's local log files. Their format isn't official, so a Claude Code update could break it for a while.
 
 ## More
 
-- 📖 [Full guide](docs/GUIDE.md): several computers, several accounts, all commands, uninstalling
-- 🐛 [Report a bug](https://github.com/us-urvin/worklog-for-claude-code/issues/new/choose) · 🔐 [Security](SECURITY.md) · 🤝 [Contributing](CONTRIBUTING.md) · 📝 [Changelog](CHANGELOG.md)
+- 📖 [Full guide](https://github.com/us-urvin/worklog-for-claude-code/blob/main/docs/GUIDE.md): several computers, several accounts, all commands, uninstalling
+- 🐛 [Report a bug](https://github.com/us-urvin/worklog-for-claude-code/issues/new/choose) · 🔐 [Security](https://github.com/us-urvin/worklog-for-claude-code/blob/main/SECURITY.md) · 🤝 [Contributing](https://github.com/us-urvin/worklog-for-claude-code/blob/main/CONTRIBUTING.md) · 📝 [Changelog](https://github.com/us-urvin/worklog-for-claude-code/blob/main/CHANGELOG.md)
 
 ## License
 
-[MIT](LICENSE) © 2026 Urvin
+[MIT](https://github.com/us-urvin/worklog-for-claude-code/blob/main/LICENSE) © 2026 Urvin

{worklog_for_claude_code-0.1.0 → worklog_for_claude_code-0.2.0}/pyproject.toml

@@ -5,14 +5,14 @@ build-backend = "setuptools.build_meta"
 [project]
 # Distribution name on PyPI. The command stays `claude-worklog` and the package stays `worklog`.
 name = "worklog-for-claude-code"
-version = "0.1.0"
-description = "Daily work log and token report across Claude Code accounts and devices, synced through your own private git repo"
+version = "0.2.0"
+description = "Daily work log and token report across Claude Code accounts and devices, plus imported claude.ai chats, synced through your own private git repo"
 readme = "README.md"
 requires-python = ">=3.10"
 license = "MIT"
 license-files = ["LICENSE"]
 authors = [{ name = "Urvin" }]
-keywords = ["claude-code", "claude", "worklog", "time-tracking", "token-usage", "developer-tools", "daily-report"]
+keywords = ["claude-code", "claude", "claude-ai", "worklog", "time-tracking", "token-usage", "developer-tools", "daily-report"]
 classifiers = [
     "Development Status :: 4 - Beta",
     "Environment :: Console",

worklog_for_claude_code-0.2.0/tests/test_chats.py

@@ -0,0 +1,407 @@
+"""claude.ai chat export import. Every export here is synthetic, built in code."""
+import http.client
+import json
+import logging
+import os
+import socket
+import subprocess
+import threading
+import time
+import zipfile
+from datetime import date, datetime, timedelta, timezone
+from pathlib import Path
+
+import pytest
+
+from worklog import chats, cli, identity, store, web
+from worklog.config import ConfigError, dump_config, load_config
+from worklog.report import aggregate, render
+
+ACCT = "66666666-aaaa-4aaa-8aaa-666666666666"
+SECRET_REPLY = "ASSISTANT-REPLY-should-never-be-stored"
+SECRET_SUMMARY = "CHAT-SUMMARY-should-never-be-stored"
+SECRET_ATTACHMENT = "ATTACHMENT-CONTENT-should-never-be-stored"
+SECRET_THINKING = "THINKING-should-never-be-stored"
+SECRET_TOOL_INPUT = "TOOL-INPUT-should-never-be-stored"
+
+
+# ------------------------------------------------------------------ helpers
+def ts(d: date, h: int, m: int = 0) -> str:
+    """A local wall-clock time as the export writes it (UTC with Z)."""
+    local = datetime(d.year, d.month, d.day, h, m).astimezone()
+    return local.astimezone(timezone.utc).isoformat().replace("+00:00", "Z")
+
+
+def human(text: str, when: str, **kw) -> dict:
+    return {"uuid": f"h-{when}-{len(text)}", "sender": "human", "text": text, "created_at": when,
+            "content": [{"type": "text", "text": text}] if text else [], "attachments": [], "files": [],
+            "parent_message_uuid": "p", **kw}
+
+
+def assistant(when: str, tools: list[dict] | None = None) -> dict:
+    content = [{"type": "thinking", "thinking": SECRET_THINKING},
+               {"type": "text", "text": SECRET_REPLY}] + (tools or [])
+    return {"uuid": f"a-{when}", "sender": "assistant", "text": SECRET_REPLY, "created_at": when,
+            "content": content, "attachments": [], "files": [], "parent_message_uuid": "p"}
+
+
+def tool(name: str, tid: str, **inp) -> dict:
+    return {"type": "tool_use", "id": tid, "name": name, "integration_name": "Claude Code",
+            "input": {"content": SECRET_TOOL_INPUT, **inp}}
+
+
+def conv(uuid: str, msgs: list[dict], name: str = "Plan the launch", updated: str = "2026-09-21T10:00:00Z") -> dict:
+    return {"uuid": uuid, "name": name, "summary": SECRET_SUMMARY, "created_at": msgs[0]["created_at"] if msgs else "",
+            "updated_at": updated, "account": {"uuid": ACCT}, "chat_messages": msgs}
+
+
+def write_zip(path: Path, convs: list[dict], member: str = "conversations.json") -> Path:
+    with zipfile.ZipFile(path, "w", zipfile.ZIP_DEFLATED) as zf:
+        zf.writestr(member, json.dumps(convs))
+    return path
+
+
+D = date(2026, 9, 20)
+
+
+@pytest.fixture
+def home(tmp_path, monkeypatch):
+    h = tmp_path / "home"
+    (h / ".claude" / "projects").mkdir(parents=True)
+    for var, sub in (("HOME", ""), ("XDG_CONFIG_HOME", ".config"), ("XDG_STATE_HOME", ".local/state"),
+                     ("XDG_DATA_HOME", ".local/share")):
+        monkeypatch.setenv(var, str(h / sub) if sub else str(h))
+    state = h / ".local/state/claude-worklog"
+    for mod in (identity, store, cli, chats):
+        monkeypatch.setattr(mod, "STATE_DIR", state)
+    assert Path.home() == h
+    return h
+
+
+@pytest.fixture
+def cfg(home):
+    repo = home / "data"
+    repo.mkdir()
+    subprocess.run(["git", "init", "-q", str(repo)], check=True)
+    p = home / ".config/claude-worklog/config.toml"
+    p.parent.mkdir(parents=True, exist_ok=True)
+    p.write_text(f'device_name = "dev1"\nstore_repo = "{repo}"\n'
+                 '[[accounts]]\nname = "personal"\nconfig_dir = "~/.claude"\n'
+                 '[[accounts]]\nname = "work"\nconfig_dir = "~/.claude-work"\n'
+                 '[collect]\nmax_prompt_chars = 40\n[report]\nuse_ai = false\n')
+    return load_config(p)
+
+
+def records(cfg) -> dict[str, dict]:
+    return {str(p.relative_to(cfg.store_repo)): json.loads(p.read_text())
+            for p in sorted(cfg.store_repo.rglob("records/**/*.json"))}
+
+
+def commits(cfg) -> int:
+    out = subprocess.run(["git", "-C", str(cfg.store_repo), "rev-list", "--count", "--all"],
+                         capture_output=True, text=True)
+    return int(out.stdout.strip() or 0)
+
+
+def basic_export() -> list[dict]:
+    return [conv("c1", [human("add a login page", ts(D, 9)), assistant(ts(D, 9, 1))])]
+
+
+# ------------------------------------------------------------------ reading
+def test_reads_zip_json_and_folder_and_merges_parts(tmp_path):
+    convs = basic_export()
+    (tmp_path / "conversations.json").write_text(json.dumps(convs))
+    assert chats.load_conversations(tmp_path / "conversations.json")[0][0]["uuid"] == "c1"
+    z = write_zip(tmp_path / "conversations-000.zip", convs, member="export/conversations.json")
+    assert chats.load_conversations(z)[0][0]["uuid"] == "c1"
+
+    folder = tmp_path / "dl"
+    folder.mkdir()
+    old = conv("c1", [human("old", ts(D, 9))], updated="2026-09-20T00:00:00Z")
+    new = conv("c1", [human("new", ts(D, 9))], updated="2026-09-22T00:00:00Z")
+    write_zip(folder / "conversations-000.zip", [new, conv("c2", [human("x", ts(D, 10))])])
+    write_zip(folder / "conversations-001.zip", [old])
+    (folder / "memories-000.zip").write_bytes(b"ignored")
+    got, bad, files = chats.load_conversations(folder)
+    assert sorted(c["uuid"] for c in got) == ["c1", "c2"] and bad == 0 and len(files) == 2
+    assert next(c for c in got if c["uuid"] == "c1")["chat_messages"][0]["text"] == "new"
+
+
+@pytest.mark.parametrize("content, msg", [
+    (json.dumps([{"id": "x", "mapping": {}}]), "ChatGPT"),
+    (json.dumps({"conversations": []}), "expected a list"),
+    ("{not json", "not valid JSON"),
+    (json.dumps([{"foo": 1}]), "no chat_messages"),
+])
+def test_rejects_files_that_are_not_claude_exports(tmp_path, content, msg):
+    p = tmp_path / "conversations.json"
+    p.write_text(content)
+    with pytest.raises(chats.ExportError, match=msg):
+        chats.load_conversations(p)
+
+
+def test_rejects_wrong_zip_and_oversized_member(tmp_path, monkeypatch):
+    with pytest.raises(chats.ExportError, match="no conversations.json"):
+        chats.load_conversations(write_zip(tmp_path / "memories-000.zip", [], member="memories.json"))
+    (tmp_path / "broken.zip").write_bytes(b"PK\x03\x04 not really")
+    with pytest.raises(chats.ExportError, match="not a readable zip"):
+        chats.load_conversations(tmp_path / "broken.zip")
+    monkeypatch.setattr(chats, "MAX_JSON_BYTES", 100)
+    with pytest.raises(chats.ExportError, match="larger than"):
+        chats.load_conversations(write_zip(tmp_path / "big.zip", [conv("c", [human("x" * 500, ts(D, 9))])]))
+    with pytest.raises(chats.ExportError, match="does not exist"):
+        chats.load_conversations(tmp_path / "missing.zip")
+
+
+def test_malformed_chats_and_messages_are_counted_not_fatal(tmp_path):
+    good = conv("c1", [human("ok", ts(D, 9)), "junk", human("no time", "yesterday"), human("x", None)])
+    p = tmp_path / "conversations.json"
+    p.write_text(json.dumps([good, {"uuid": "c2"}, 7]))
+    convs, bad, _ = chats.load_conversations(p)
+    assert len(convs) == 1 and bad == 2
+    days, bad_msgs = chats.split_days(convs[0])
+    assert bad_msgs == 3 and days[D].prompts == ["ok"]
+
+
+# ------------------------------------------------------------------ per-day rules
+def test_only_real_prompts_are_kept_and_days_split_at_local_midnight():
+    msgs = [
+        human("first question", ts(D, 23, 50)),
+        human("", ts(D, 23, 51)),  # tool-loop turn
+        human("", ts(D, 23, 52), content=[{"type": "injected_prompt_block", "prompt": "system stuff"}]),
+        human("first question", ts(D, 23, 55)),  # edited/retried branch repeats the prompt
+        assistant(ts(D, 23, 56)),
+        human("after midnight", ts(D + timedelta(days=1), 0, 10)),
+    ]
+    days, bad = chats.split_days(conv("c1", msgs))
+    assert bad == 0 and set(days) == {D, D + timedelta(days=1)}
+    assert days[D].prompts == ["first question"] and days[D].messages == 5
+    assert days[D + timedelta(days=1)].prompts == ["after midnight"]
+    since, _ = chats.split_days(conv("c1", msgs), since=D + timedelta(days=1))
+    assert set(since) == {D + timedelta(days=1)}
+
+
+def test_record_redacts_truncates_and_never_holds_claudes_text(cfg):
+    key = "sk-ant-api03-" + "a" * 30
+    msgs = [human(f"my key is {key} and this prompt is far longer than forty characters", ts(D, 9),
+                  attachments=[{"file_name": "notes.txt", "extracted_content": SECRET_ATTACHMENT}]),
+            assistant(ts(D, 9, 5), [tool("Write", "toolu_01A", file_path="/home/claude/app/src/main.py")])]
+    days, _ = chats.split_days(conv("c1", msgs, name=f"Debug {key}"))
+    rec = chats.build_record(days[D], "personal", cfg)
+    blob = json.dumps(rec)
+    for secret in (key, SECRET_REPLY, SECRET_SUMMARY, SECRET_ATTACHMENT, SECRET_THINKING, SECRET_TOOL_INPUT,
+                   "toolu_01A", ACCT):
+        assert secret not in blob
+    assert rec["prompts"][0].startswith("my key is [REDACTED]") and len(rec["prompts"][0]) == 40
+    assert rec["title"] == "Debug [REDACTED]"
+    assert rec["source"] == "claude.ai" and rec["session_id"] == "chat-c1"
+    assert rec["tokens"] == {"input": 0, "output": 0, "cache_creation": 0, "cache_read": 0}
+    assert rec["project"]["name"] == "app" and rec["files_touched"] == ["src/main.py"]
+    assert rec["tool_calls"] == {"Write": 1}
+
+
+def test_untitled_chat_and_assistant_only_day():
+    days, _ = chats.split_days(conv("c1", [human("hi", ts(D, 9))], name=""))
+    assert chats.build_record(days[D], "personal", load_cfg_stub())["title"] == "Untitled chat"
+    days, _ = chats.split_days(conv("c2", [assistant(ts(D, 9))]))
+    assert chats.build_record(days[D], "personal", load_cfg_stub()) is None
+
+
+def load_cfg_stub():
+    from worklog.config import Account, Config
+    return Config(device="dev1", store_repo=Path("/nonexistent"), accounts=[Account("personal", Path("/x"))])
+
+
+# ------------------------------------------------------------------ import
+def test_import_is_idempotent_and_an_older_export_never_overwrites(cfg, tmp_path):
+    newer = conv("c1", [human("one", ts(D, 9)), human("two", ts(D, 10))], updated="2026-09-22T00:00:00Z")
+    res = chats.import_chats(cfg, write_zip(tmp_path / "conversations-000.zip", [newer]))
+    assert (res.chats, res.days, res.written, res.account) == (1, 1, 1, "personal")
+    path = f"records/{D.isoformat()}/dev1/personal/chat-c1.json"
+    assert records(cfg)[path]["prompts"] == ["one", "two"] and commits(cfg) == 1
+
+    again = chats.import_chats(cfg, tmp_path / "conversations-000.zip")
+    assert again.written == 0 and commits(cfg) == 1
+
+    older = conv("c1", [human("one", ts(D, 9))], updated="2026-09-21T00:00:00Z")
+    res = chats.import_chats(cfg, write_zip(tmp_path / "old.zip", [older]))
+    assert res.stale == 1 and res.written == 0 and records(cfg)[path]["prompts"] == ["one", "two"]
+    assert chats.load_state()["last_import"]["chats"] == 1
+
+
+def test_chats_that_are_local_claude_code_sessions_are_skipped(cfg, home, tmp_path):
+    d = home / ".claude/projects/-work"
+    d.mkdir(parents=True)
+    (d / "s1.jsonl").write_text(json.dumps({"type": "assistant", "timestamp": ts(D, 9), "message": {
+        "content": [{"type": "tool_use", "id": "toolu_01LOCAL", "name": "Bash", "input": {}}]}}) + "\n")
+    os.utime(d / "s1.jsonl", (time.time(), time.time()))
+    dup = conv("c1", [human("fix it", ts(D, 9)), assistant(ts(D, 9, 1), [tool("Bash", "toolu_01LOCAL")])])
+    other = conv("c2", [human("other", ts(D, 9)), assistant(ts(D, 9, 1), [tool("Bash", "toolu_01CLOUD")])])
+    res = chats.import_chats(cfg, write_zip(tmp_path / "c.zip", [dup, other]))
+    assert res.duplicates == 1 and res.days == 1
+    assert list(records(cfg)) == [f"records/{D.isoformat()}/dev1/personal/chat-c2.json"]
+
+
+def test_account_choice_is_remembered_and_records_move(cfg, tmp_path):
+    z = write_zip(tmp_path / "c.zip", basic_export())
+    assert chats.import_chats(cfg, z).account == "personal"
+    with pytest.raises(chats.ExportError, match="no account named"):
+        chats.import_chats(cfg, z, account="nope")
+    res = chats.import_chats(cfg, z, account="work")
+    assert res.account == "work" and list(records(cfg)) == [f"records/{D.isoformat()}/dev1/work/chat-c1.json"]
+    assert chats.import_chats(cfg, z).account == "work"  # remembered for this claude.ai login
+    state = json.loads(chats.state_path().read_text())
+    assert ACCT not in json.dumps(state) and oct(chats.state_path().stat().st_mode)[-3:] == "600"
+
+
+def test_cli_import_chats(cfg, tmp_path, capsys, monkeypatch):
+    monkeypatch.setattr(cli, "load_config", lambda: cfg)
+    z = write_zip(tmp_path / "c.zip", basic_export())
+    try:
+        assert cli.main(["import-chats", str(z)]) == 0
+        assert "Imported 1 chat(s)" in capsys.readouterr().out
+        (tmp_path / "bad.json").write_text("[]x")
+        assert cli.main(["import-chats", str(tmp_path / "bad.json")]) == 1
+        assert "not valid JSON" in capsys.readouterr().out
+    finally:
+        root = logging.getLogger()
+        for h in root.handlers[:]:  # cli._setup_logging adds handlers on every call
+            root.removeHandler(h)
+            h.close()
+
+
+# ------------------------------------------------------------------ downloads
+def test_downloads_scan_is_opt_in_and_skips_partial_recent_and_seen(cfg, home):
+    dl = home / "Downloads"
+    dl.mkdir()
+    z = write_zip(dl / "conversations-000.zip", basic_export())
+    old = time.time() - 60
+    os.utime(z, (old, old))
+    assert chats.scan_downloads(cfg) == []  # toggle is off: nothing is read
+
+    cfg.watch_downloads = True
+    (dl / "conversations-000.zip.crdownload").write_bytes(b"")
+    assert chats.scan_downloads(cfg) == []  # browser still writing
+    (dl / "conversations-000.zip.crdownload").unlink()
+    fresh = write_zip(dl / "conversations-001.zip", [conv("c9", [human("new", ts(D, 9))])])
+    assert chats.pending_downloads(cfg) == [z]  # fresh file has not settled yet
+    results = chats.scan_downloads(cfg)
+    assert len(results) == 1 and results[0].days == 1
+    assert chats.scan_downloads(cfg) == []  # already imported
+    assert z.exists() and fresh.exists()  # never moved or deleted
+
+    bad = dl / "conversations-x.json"
+    bad.write_text(json.dumps([{"mapping": {}}]))
+    os.utime(bad, (old, old))
+    os.utime(fresh, (old, old))
+    results = chats.scan_downloads(cfg)
+    assert len(results) == 1  # the bad file is skipped, the settled one imported
+    assert chats.pending_downloads(cfg) == []  # and the bad one is not retried every sync
+
+
+def test_downloads_dir_from_xdg_user_dirs(home):
+    from worklog.config import default_downloads_dir
+    assert default_downloads_dir() == home / "Downloads"
+    (home / ".config").mkdir(exist_ok=True)
+    (home / ".config/user-dirs.dirs").write_text('XDG_DOWNLOAD_DIR="$HOME/Téléchargements"\n')
+    assert default_downloads_dir() == home / "Téléchargements"
+
+
+# ------------------------------------------------------------------ report + config
+def test_report_counts_chats_once_across_devices_without_tokens(cfg, tmp_path):
+    chats.import_chats(cfg, write_zip(tmp_path / "c.zip", basic_export()))
+    rec = next(iter(records(cfg).values()))
+    copy = dict(rec, device="laptop")
+    accounts, projects = aggregate([rec, copy])
+    p = projects["chat:claude.ai chat"]
+    assert p.chats == 1 and p.sessions == 0 and p.titles == ["Plan the launch"]
+    text = render(D, cfg, accounts, projects, {})
+    assert "1 claude.ai chat(s)" in text and "no token data" in text and "Token usage" not in text
+    assert "- Plan the launch" in text
+
+    body, err = web.page_day(web.App(8765), cfg, D)
+    assert err is None and "1 claude.ai chat" in body and "no data from claude.ai" in body
+
+
+def test_config_chat_section_roundtrip_and_validation(cfg):
+    cfg.chat_account, cfg.watch_downloads, cfg.downloads_dir = "work", True, Path.home() / "dl"
+    p = cfg.path.with_name("roundtrip.toml")
+    p.write_text(dump_config(cfg))
+    again = load_config(p)
+    assert (again.chats_account, again.watch_downloads, again.downloads) == ("work", True, Path.home() / "dl")
+    p.write_text(dump_config(cfg).replace('account = "work"', 'account = "ghost"'))
+    with pytest.raises(ConfigError, match="ghost"):
+        load_config(p)
+    assert load_config(cfg.path).chats_account == "personal"  # default: first account
+
+
+# ------------------------------------------------------------------ web
+@pytest.fixture
+def server(cfg, monkeypatch):
+    monkeypatch.setattr(web, "load_config", lambda: cfg)
+    with socket.socket() as s:
+        s.bind(("127.0.0.1", 0))
+        port = s.getsockname()[1]
+    app = web.App(port)
+    srv = web._Server(("127.0.0.1", port), web.make_handler(app))
+    threading.Thread(target=srv.serve_forever, daemon=True).start()
+    yield app, port
+    srv.shutdown()
+    srv.server_close()
+
+
+def _post(port, path, body: bytes, headers: dict):
+    c = http.client.HTTPConnection("127.0.0.1", port, timeout=10)
+    c.request("POST", path, body=body, headers={"Host": f"localhost:{port}", **headers})
+    r = c.getresponse()
+    out = r.status, r.read()
+    c.close()
+    return out
+
+
+def test_upload_needs_csrf_and_imports_in_the_background(server, cfg, tmp_path):
+    app, port = server
+    data = write_zip(tmp_path / "c.zip", basic_export()).read_bytes()
+    assert _post(port, "/sources/upload", data, {"X-CSRF-Token": "wrong"})[0] == 403
+    assert _post(port, "/sources/upload", data, {"X-CSRF-Token": app.csrf,
+                                                  "Origin": "http://evil.example"})[0] == 403
+    status, body = _post(port, "/sources/upload", data, {"X-CSRF-Token": app.csrf,
+                                                          "X-File-Name": "conversations-000.zip"})
+    assert status == 202, body
+    for _ in range(100):
+        if not app.jobs.snapshot()["running"] and app.jobs.snapshot()["finished"]:
+            break
+        time.sleep(0.05)
+    assert "Imported 1 chat(s)" in app.jobs.snapshot()["message"]
+    assert list((chats.STATE_DIR / "uploads").iterdir()) == []  # temp copy removed
+    assert any(k.endswith("chat-c1.json") for k in records(cfg))
+
+
+def test_upload_rejects_oversize_and_bad_files(server, monkeypatch):
+    app, port = server
+    monkeypatch.setattr(web, "MAX_UPLOAD", 10)
+    assert _post(port, "/sources/upload", b"x" * 11, {"X-CSRF-Token": app.csrf})[0] == 413
+    monkeypatch.setattr(web, "MAX_UPLOAD", 1000)
+    status, _ = _post(port, "/sources/upload", b'[{"mapping": {}}]', {"X-CSRF-Token": app.csrf})
+    assert status == 202
+    for _ in range(100):
+        if app.jobs.snapshot()["finished"]:
+            break
+        time.sleep(0.05)
+    snap = app.jobs.snapshot()
+    assert not snap["ok"] and "ChatGPT" in snap["message"]
+
+
+def test_sources_page_and_settings(server, cfg, monkeypatch):
+    app, _ = server
+    html_out = web.page_sources(app, cfg)
+    assert "claude.ai chats" in html_out and "Not imported yet" in html_out and "Coming later" in html_out
+    assert "Nothing in your Downloads folder is read" in html_out
+    saved = []
+    monkeypatch.setattr(web, "save_config", lambda c: saved.append(c))
+    loc, err, _ = app.post("/sources/settings", {"account": "work", "watch": "1", "downloads": "~/dl"})
+    assert err is None and loc == "/sources?ok=chat_saved"
+    new = saved[-1]
+    assert new.chat_account == "work" and new.watch_downloads and new.downloads_dir == Path.home() / "dl"
+    assert app.post("/sources/settings", {"account": "work", "downloads": "/etc"})[1]
+    assert app.post("/sources/settings", {"account": "ghost"})[1]

{worklog_for_claude_code-0.1.0 → worklog_for_claude_code-0.2.0}/worklog/__init__.py

@@ -1,3 +1,3 @@
 """claude-worklog: daily work log + token report for Claude Code."""
 
-__version__ = "0.1.0"
+__version__ = "0.2.0"