@cerefox/memory 0.4.3 → 0.5.0

package/docs/guides/cli.md

@@ -0,0 +1,481 @@
+# Cerefox CLI Reference
+
+Comprehensive reference for every `cerefox` subcommand. For tutorials and walkthroughs see [`quickstart.md`](quickstart.md); for the agent-via-CLI use case (Claude Code, Codex CLI, opencode, OpenClaw, Hermes, …) see [Path C in `connect-agents.md`](connect-agents.md#path-c--shell-cli-for-local-coding-agents).
+
+> `--help` is canonical. If anything in this document disagrees with `cerefox <subcommand> --help`, trust `--help` and file an issue against this guide.
+
+## Setup
+
+Every command reads configuration from `.env` in the working directory (or environment variables — see [`configuration.md`](configuration.md)). Required at minimum:
+
+- `CEREFOX_SUPABASE_URL` and `CEREFOX_SUPABASE_KEY` for any command that talks to Supabase
+- `OPENAI_API_KEY` (or `CEREFOX_FIREWORKS_API_KEY`) for any command that embeds (ingest, search)
+- `CEREFOX_DATABASE_URL` only for the `scripts/db_*.py` deployment scripts (not the `cerefox` CLI itself)
+
+Invoke any command with `uv run cerefox <subcommand>`. Inside an activated venv, `cerefox <subcommand>` works too — but `uv run` is preferred (no venv activation needed; see Decision Log Q2 lesson on `uv` installation).
+
+## Commands
+
+### `cerefox ingest`
+
+**Purpose**: ingest a markdown / PDF / DOCX file (or stdin) into the knowledge base.
+
+**Synopsis**:
+```
+cerefox ingest [OPTIONS] [PATH]
+cerefox ingest --paste --title "<title>" [OPTIONS]   # stdin
+```
+
+**Options**:
+
+> **Flag naming**: every flag below matches its MCP-tool parameter name (e.g. `project_name` → `--project-name`). Short forms (`--project`, `-p`) remain as aliases — the long form is the canonical name shown in `--help`.
+
+| Flag (canonical) | Aliases | Type | Default | Description |
+|---|---|---|---|---|
+| `--title` | `-t` | str | filename stem | Document title. Required with `--paste`. |
+| `--project-name` | `--project`, `-p` | str | _none_ | Project name to assign the document to (created if missing). |
+| `--paste` | — | flag | off | Read markdown from stdin. Requires `--title`. |
+| `--metadata` | `-m` | JSON | `{}` | Extra metadata as a JSON object, e.g. `'{"tags":["work"]}'`. |
+| `--update-if-exists` | `--update` | flag | off | Title/source-path-based fallback update. Mutually exclusive with `--document-id`. |
+| `--document-id` | — | UUID | _none_ | Deterministic ID-based update. Errors if the document doesn't exist. |
+| `--source` | — | str | `paste` / `file` | Source label recorded on the document. |
+| `--author` | — | str | `CEREFOX_AUTHOR_NAME` or `unknown` | Audit-log author identity. |
+| `--author-type` | — | `user`\|`agent` | `CEREFOX_AUTHOR_TYPE` or `user` | Caller type. Agent writes auto-routed to `pending_review`. |
+
+**Examples**:
+```bash
+# Minimal: ingest a file
+cerefox ingest notes.md
+
+# Paste from stdin
+printf '# Title\n\nbody' | cerefox ingest --paste --title "Title"
+
+# Agent ingestion with full attribution
+cerefox ingest notes.md \
+  --author "claude-code" --author-type "agent" \
+  --project-name "research" --metadata '{"type":"design-doc"}'
+
+# Deterministic update (preferred — agents should search → grab ID → ingest)
+cerefox ingest --paste --title "Same Title" \
+  --document-id "abc12345-..." \
+  --author "claude-code" --author-type "agent"
+```
+
+**Output**: human-readable summary line(s) — "Ingested" or "Updated" with the document ID, chunk count, character count.
+
+**Exit codes**: `0` success, `1` on validation error (missing `--title`, invalid JSON, document-not-found, mutually-exclusive flags, etc.).
+
+**MCP equivalent**: [`cerefox_ingest`](../../AGENT_GUIDE.md).
+
+---
+
+### `cerefox ingest-dir`
+
+**Purpose**: bulk-ingest every matching file in a directory.
+
+**Synopsis**:
+```
+cerefox ingest-dir [OPTIONS] DIRECTORY
+```
+
+**Options**:
+
+| Flag | Type | Default | Description |
+|---|---|---|---|
+| `--pattern TEXT` | glob | `*.md` | Glob pattern. Examples: `**/*.md`, `*.pdf`. |
+| `--project-name TEXT` (alias: `--project`, `-p`) | str | _none_ | Project to assign every document to. |
+| `--recursive / --no-recursive` | flag | `--no-recursive` | Recurse into sub-directories. |
+| `--dry-run` | flag | off | Print files that would be ingested; do nothing. |
+| `--update-if-exists` (alias: `--update`) | flag | off | Update existing documents by source path. |
+| `-m, --metadata TEXT` | JSON | `{}` | JSON metadata applied to every file in the run. |
+| `--author TEXT` | str | `CEREFOX_AUTHOR_NAME` or `unknown` | Audit-log author identity (applies to every write). |
+| `--author-type [user\|agent]` | choice | `CEREFOX_AUTHOR_TYPE` or `user` | Caller type. |
+
+**Examples**:
+```bash
+# Bulk import research notes with shared metadata
+cerefox ingest-dir ./research-notes --recursive \
+  --project-name "research" --metadata '{"type":"research","status":"active"}'
+
+# Re-ingest after editing files
+cerefox ingest-dir ./notes --update-if-exists
+```
+
+**Output**: one line per file showing `✓` (ingested), `↑` (updated), `⏭` (skipped: hash match), or `❌` (error); summary at the end.
+
+**Exit codes**: `0` even if some files errored — per-file errors are counted but the command itself does not fail unless argument validation fails.
+
+---
+
+### `cerefox search`
+
+**Purpose**: search the knowledge base (hybrid FTS + semantic by default).
+
+**Synopsis**:
+```
+cerefox search [OPTIONS] QUERY
+```
+
+**Options**:
+
+| Flag | Type | Default | Description |
+|---|---|---|---|
+| `-m, --mode [hybrid\|fts\|semantic]` | choice | `hybrid` | Search mode. |
+| `--match-count INTEGER` (alias: `--count`, `-n`) | int | `10` | Number of results. |
+| `--project-name TEXT` (alias: `--project`, `-p`) | str | _none_ | Limit to a project by name. |
+| `--alpha FLOAT` | float | `0.7` | FTS/semantic weight (hybrid only). |
+| `--min-score FLOAT` | float | `CEREFOX_MIN_SEARCH_SCORE` or `0.50` | Minimum cosine similarity (hybrid/semantic only). |
+| `--metadata-filter TEXT` (alias: `--filter`, `-f`) | JSON | _none_ | JSONB metadata containment filter, e.g. `'{"type":"decision"}'`. |
+| `--requestor TEXT` | str | `CEREFOX_REQUESTOR_NAME` or `user` | Identity recorded in the usage log. |
+
+**Examples**:
+```bash
+cerefox search "OAuth design"
+cerefox search "decisions" --metadata-filter '{"type":"decision-log"}' --match-count 5
+cerefox search "what we tried" --mode semantic --requestor "claude-code"
+```
+
+**Output**: numbered result list with title, score, and 300-char preview per hit. Final line shows total results + bytes.
+
+**Exit codes**: `0` on success. Note: as of v0.1.17, the CLI logs usage in a try/except so a usage-logging error does not affect the user-visible output (closes the failure mode that produced cerefox#27).
+
+**MCP equivalent**: [`cerefox_search`](../../AGENT_GUIDE.md).
+
+---
+
+### `cerefox get-doc`
+
+**Purpose**: print the full markdown content of a document to stdout.
+
+**Synopsis**:
+```
+cerefox get-doc [OPTIONS] DOCUMENT_ID
+```
+
+**Options**:
+
+| Flag | Type | Default | Description |
+|---|---|---|---|
+| `--version-id TEXT` (alias: `--version`) | UUID | _none_ (current) | Archived version UUID — get from `cerefox list-versions`. |
+| `--requestor TEXT` | str | `CEREFOX_REQUESTOR_NAME` or `user` | Identity recorded in the usage log. |
+
+**Examples**:
+```bash
+cerefox get-doc abc12345-...
+cerefox get-doc abc12345-... --version-id <version-uuid>     # archived
+cerefox get-doc abc12345-... | bat -l md                  # pipe to viewer
+```
+
+**Output**: title + metadata line, blank line, then raw markdown.
+
+**MCP equivalent**: [`cerefox_get_document`](../../AGENT_GUIDE.md).
+
+---
+
+### `cerefox list-docs`
+
+**Purpose**: list documents in the knowledge base.
+
+**Synopsis**:
+```
+cerefox list-docs [OPTIONS]
+```
+
+**Options**:
+
+| Flag | Type | Default | Description |
+|---|---|---|---|
+| `--project-name TEXT` (alias: `--project`, `-p`) | str | _none_ | Filter by project ID or name. |
+| `-n, --limit INTEGER` | int | `20` | Max rows. |
+
+**Output**: tabular `id | chunk_count | total_chars | title` listing. CLI-only — there is no MCP equivalent.
+
+---
+
+### `cerefox list-versions`
+
+**Purpose**: list all archived versions of a document.
+
+**Synopsis**:
+```
+cerefox list-versions [OPTIONS] DOCUMENT_ID
+```
+
+**Options**:
+
+| Flag | Type | Default | Description |
+|---|---|---|---|
+| `--requestor TEXT` | str | `CEREFOX_REQUESTOR_NAME` or `user` | Identity recorded in the usage log. |
+
+**Output**: table with version number, created timestamp, source, chunk/char counts, and version UUID. Pass the UUID to `cerefox get-doc --version-id <uuid>` to retrieve the archived content.
+
+**MCP equivalent**: [`cerefox_list_versions`](../../AGENT_GUIDE.md).
+
+---
+
+### `cerefox list-projects`
+
+**Purpose**: list all projects.
+
+**Synopsis**:
+```
+cerefox list-projects [OPTIONS]
+```
+
+**Options**:
+
+| Flag | Type | Default | Description |
+|---|---|---|---|
+| `--requestor TEXT` | str | `CEREFOX_REQUESTOR_NAME` or `user` | Identity recorded in the usage log. |
+
+**MCP equivalent**: [`cerefox_list_projects`](../../AGENT_GUIDE.md).
+
+---
+
+### `cerefox list-metadata-keys`
+
+**Purpose**: discover metadata keys used across all documents (with example values and document counts).
+
+**Synopsis**: `cerefox list-metadata-keys`
+
+**MCP equivalent**: [`cerefox_list_metadata_keys`](../../AGENT_GUIDE.md).
+
+---
+
+### `cerefox metadata-search`
+
+**Purpose**: find documents by metadata key-value criteria (no text query needed).
+
+**Synopsis**:
+```
+cerefox metadata-search --metadata-filter '<json>' [OPTIONS]
+```
+
+**Options**:
+
+| Flag | Type | Default | Description |
+|---|---|---|---|
+| `--metadata-filter TEXT` (alias: `--filter`) | JSON | **required** | Metadata filter, e.g. `'{"type":"decision-log"}'`. |
+| `--project-name TEXT` (alias: `--project`) | str | _none_ | Filter by project name. |
+| `--updated-since TEXT` | ISO-8601 | _none_ | Documents updated after this timestamp. |
+| `--created-since TEXT` | ISO-8601 | _none_ | Documents created after this timestamp. |
+| `--limit INTEGER` | int | `10` | Max results. |
+| `--include-content` | flag | off | Include full document content (slower; subject to byte budget). |
+| `--requestor TEXT` | str | `CEREFOX_REQUESTOR_NAME` or `user` | Identity recorded in the usage log. |
+
+**Examples**:
+```bash
+cerefox metadata-search --metadata-filter '{"type":"decision-log"}' --updated-since 2026-05-01
+cerefox metadata-search --metadata-filter '{"status":"active"}' --project-name "research" --include-content
+```
+
+**MCP equivalent**: [`cerefox_metadata_search`](../../AGENT_GUIDE.md).
+
+---
+
+### `cerefox get-audit-log`
+
+**Purpose**: query the immutable audit log (who changed what, when).
+
+**Synopsis**:
+```
+cerefox get-audit-log [OPTIONS]
+```
+
+**Options**:
+
+| Flag | Type | Default | Description |
+|---|---|---|---|
+| `--document-id TEXT` | UUID | _none_ | Filter to a single document. |
+| `--author TEXT` | str | _none_ | Filter by author name (exact match). |
+| `--operation [create\|update-content\|update-metadata\|delete\|status-change\|archive\|unarchive\|restore]` | choice | _none_ | Filter by operation type. |
+| `--since TEXT` | ISO-8601 | _none_ | Lower bound on `created_at`. |
+| `--until TEXT` | ISO-8601 | _none_ | Upper bound on `created_at`. |
+| `--limit INTEGER` | int | `50` | Max rows. |
+| `--json` | flag | off | Emit one JSON object per line (for piping to `jq` / scripts). |
+| `--requestor TEXT` | str | `CEREFOX_REQUESTOR_NAME` or `user` | Identity recorded in the usage log. |
+
+**Examples**:
+```bash
+# All audit entries in the last week
+cerefox get-audit-log --since 2026-05-11
+
+# All edits by a specific agent
+cerefox get-audit-log --author "claude-code" --operation update-content
+
+# JSON output, piped to jq
+cerefox get-audit-log --json --limit 1000 | jq 'select(.author_type == "agent")'
+```
+
+**MCP equivalent**: [`cerefox_get_audit_log`](../../AGENT_GUIDE.md).
+
+---
+
+### `cerefox delete-doc`
+
+**Purpose**: **soft-delete** a document — moves it to trash, recoverable. The CLI cannot permanently delete or restore; see [Destructive operations and the trust model](access-paths.md#destructive-operations-and-the-trust-model) for the rationale.
+
+**Synopsis**: `cerefox delete-doc [OPTIONS] DOCUMENT_ID`
+
+**Options**:
+
+| Flag | Type | Default | Description |
+|---|---|---|---|
+| `-y, --yes` | flag | off | Skip confirmation prompt. Required for non-interactive use (agents, scripts). |
+| `--author` | str | `CEREFOX_AUTHOR_NAME` or `unknown` | Identity recorded in the audit log. |
+| `--author-type` | `user`\|`agent` | `CEREFOX_AUTHOR_TYPE` or `user` | Caller type, recorded in the audit log. |
+
+**What this command does:**
+- Sets `deleted_at` on the document row. The document stays in the database.
+- Excludes the document from search and from `cerefox list-docs`.
+- Writes an immutable `delete` audit-log entry with the resolved author / author_type and timestamp.
+
+**What this command does NOT do:**
+- Does NOT permanently delete the document.
+- Does NOT free database storage.
+- Versions, chunks, and audit entries remain intact under the trash.
+
+**Recovery**: a soft-deleted document can be restored OR permanently purged **only from the Cerefox web UI** (Trash view). These destructive / restorative actions are intentionally web-UI-only to require human-in-the-loop confirmation. See [`access-paths.md` → Destructive operations and the trust model](access-paths.md#destructive-operations-and-the-trust-model).
+
+**Agent usage**:
+```bash
+# Required: --yes (no TTY for confirmation) + identity flags
+cerefox delete-doc <doc-id> --yes \
+  --author "claude-code" --author-type "agent"
+```
+The success message echoes the resolved author / author_type back so you can surface it to the user in your response.
+
+**Exit codes**: `0` on success. `1` on confirmation abort, missing document, or auth failure.
+
+---
+
+### `cerefox reindex`
+
+**Purpose**: re-embed chunks (e.g. after switching embedding models or pulling a schema change like title-boosting).
+
+**Synopsis**: `cerefox reindex [OPTIONS]`
+
+**Options**:
+
+| Flag | Type | Default | Description |
+|---|---|---|---|
+| `--batch INTEGER` | int | `100` | Chunks per batch. |
+| `--all` | flag | off | Reindex every chunk regardless of embedder. |
+| `--dry-run` | flag | off | Count what would be re-embedded; do nothing. |
+
+---
+
+### `cerefox config-get` / `cerefox config-set`
+
+**Purpose**: read/write runtime config in `cerefox_config` (e.g. `usage_tracking_enabled`).
+
+**Synopsis**:
+```
+cerefox config-get KEY
+cerefox config-set KEY VALUE
+```
+
+Used for toggling features at runtime without a redeploy — see [Decision Log Q1 Part 2 — usage tracking opt-in](https://github.com/fstamatelopoulos/cerefox) entry.
+
+---
+
+### `cerefox web`
+
+**Purpose**: start the FastAPI web UI (React SPA + JSON API).
+
+**Synopsis**: `cerefox web [OPTIONS]`
+
+**Options**:
+
+| Flag | Type | Default | Description |
+|---|---|---|---|
+| `--host TEXT` | str | `127.0.0.1` | Bind address. |
+| `--port INTEGER` | int | `8000` | Listen port. |
+| `--reload` | flag | off | Auto-reload on source changes (dev). |
+
+Requires the frontend to be built: `cd frontend && npm install && npm run build`.
+
+---
+
+### `cerefox mcp`
+
+**Purpose**: run the local MCP stdio server. Used by Claude Desktop / Cursor / Claude Code as a subprocess (Path A-Local in `connect-agents.md`).
+
+**Synopsis**: `cerefox mcp`
+
+No options. See [`connect-agents.md` → Path A-Local](connect-agents.md#path-a-local--local-mcp-server-cerefox-mcp).
+
+---
+
+## Environment variables
+
+The CLI reads its own runtime config from environment (or `.env`). See [`configuration.md`](configuration.md) for the full list. Most relevant to CLI behaviour:
+
+| Variable | Default | Effect |
+|---|---|---|
+| `CEREFOX_AUTHOR_NAME` | `unknown` | Default for `--author` on `ingest` / `ingest-dir`. |
+| `CEREFOX_AUTHOR_TYPE` | `user` | Default for `--author-type`. |
+| `CEREFOX_REQUESTOR_NAME` | `user` | Default for `--requestor` on read commands. |
+
+Precedence: **CLI flag > env var > built-in default**.
+
+## Exit codes
+
+| Code | Meaning |
+|---|---|
+| `0` | Success |
+| `1` | Validation error, missing config, document-not-found, etc. — see error message |
+| `2` | Click argument-parsing error (invalid Choice value, missing required arg) |
+
+## MCP tool ↔ CLI command mapping
+
+Full table for Path C agents (see [`connect-agents.md`](connect-agents.md#path-c--shell-cli-for-local-coding-agents) for context).
+
+Every MCP parameter has an exact-name CLI flag (kebab-cased). Short forms exist as aliases.
+
+| MCP tool | CLI command |
+|---|---|
+| `cerefox_search(query, match_count, project_name, metadata_filter, requestor)` | `cerefox search "<q>" --match-count N --project-name <name> --metadata-filter '<json>' --requestor <name>` |
+| `cerefox_ingest(title, content, project_name, metadata, update_if_exists, document_id, source, author, author_type)` (file) | `cerefox ingest <path> --title <t> --project-name <n> --metadata '<json>' --update-if-exists\|--document-id <uuid> --source <s> --author <a> --author-type <t>` |
+| `cerefox_ingest(...)` (paste) | `printf '...' \| cerefox ingest --paste --title "<t>"` (same flags) |
+| `cerefox_get_document(document_id, version_id, requestor)` | `cerefox get-doc <id> --version-id <vid> --requestor <name>` |
+| `cerefox_list_versions(document_id, requestor)` | `cerefox list-versions <id> --requestor <name>` |
+| `cerefox_list_projects(requestor)` | `cerefox list-projects --requestor <name>` |
+| `cerefox_list_metadata_keys()` | `cerefox list-metadata-keys` |
+| `cerefox_metadata_search(metadata_filter, project_name, updated_since, created_since, limit, include_content, requestor)` | `cerefox metadata-search --metadata-filter '<json>' --project-name <n> --updated-since <iso> --created-since <iso> --limit N --include-content --requestor <name>` |
+| `cerefox_get_audit_log(document_id, author, operation, since, until, limit, requestor)` | `cerefox get-audit-log --document-id <id> --author <a> --operation <op> --since <iso> --until <iso> --limit N --requestor <name>` |
+
+## Known issues
+
+None outstanding as of v0.1.17 (cerefox#27 — the `cerefox search` NameError — is resolved). When new bugs surface, they are tracked in the GitHub issues list; check there before relying on a behaviour the docs imply.
+
+## Common recipes
+
+### Bulk-import a directory with shared metadata
+```bash
+cerefox ingest-dir ./papers --recursive --pattern '*.pdf' \
+  --project-name "literature" \
+  --metadata '{"type":"paper","status":"reviewed"}'
+```
+
+### Update a document by ID (preferred pattern for agents)
+```bash
+# Step 1: find it
+cerefox search "the OAuth design doc" --match-count 1
+
+# Step 2: copy the id from `Doc: ... (id: <uuid>)` line
+# Step 3: update in place
+printf '%s' "$NEW_CONTENT" | cerefox ingest --paste \
+  --title "OAuth 2.1 Design Document" \
+  --document-id "<uuid>" \
+  --author "claude-code" --author-type "agent"
+```
+
+### Unattended sync job
+```bash
+# In a cron job / launchd plist. Set CEREFOX_AUTHOR_NAME=sync-script in env.
+cd /path/to/cerefox && uv run cerefox ingest-dir ~/notes --recursive --update-if-exists
+```
+
+### Use the CLI from an agent's Bash tool
+See [`connect-agents.md` → Path C](connect-agents.md#path-c--shell-cli-for-local-coding-agents) for the full setup and [`AGENT_GUIDE.md` → Using Cerefox via the CLI](../../AGENT_GUIDE.md) for the agent-facing conventions.