databar 2.0.7__tar.gz → 2.0.9__tar.gz

{databar-2.0.7/src/databar.egg-info → databar-2.0.9}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: databar
-Version: 2.0.7
+Version: 2.0.9
 Summary: Official Databar.ai Python SDK and CLI — connect to enrichments, waterfalls, and tables via api.databar.ai
 Author-email: "Databar.ai Team" <info@databar.ai>
 License: MIT License

{databar-2.0.7 → databar-2.0.9}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "databar"
-version = "2.0.7"
+version = "2.0.9"
 description = "Official Databar.ai Python SDK and CLI — connect to enrichments, waterfalls, and tables via api.databar.ai"
 readme = "README.md"
 license = { file = "LICENSE" }

{databar-2.0.7 → databar-2.0.9}/src/databar/__init__.py

@@ -58,7 +58,7 @@ from .models import (
     WaterfallEnrichment,
 )
 
-__version__ = "2.0.7"
+__version__ = "2.0.9"
 __all__ = [
     "DatabarClient",
     # exceptions

databar-2.0.9/src/databar/cli/_guide.py

@@ -0,0 +1,219 @@
+"""Embedded agent guide — printed by `databar agent-guide`."""
+
+AGENT_GUIDE = r"""# Databar — Agent Guide
+
+Databar is a data enrichment platform. Given an input (email, LinkedIn URL, company
+domain, etc.) it returns enriched data from dozens of providers. The `databar` package
+ships three interfaces: a **CLI**, a **Python SDK**, and an **MCP server**.
+
+Databar supports two enrichment workflows:
+
+- **Direct** — submit inputs and get results back immediately. Best for one-off lookups
+  or bulk CSV processing where you just need the output.
+- **Table-based** — create a table, upload rows, attach enrichments, run them. Results
+  appear as new columns on each row. Best when you want observability (see results in
+  the Databar UI), re-run enrichments on the same data, or build a persistent dataset.
+
+---
+
+## DECISION MATRIX — read this first
+
++----------------------------------+--------------------+
+| Situation                        | Use                |
++----------------------------------+--------------------+
+| MCP tools available this session | MCP (best option)  |
+| Single operation, no branching   | CLI                |
+| Conditional logic / loops /      | Python SDK         |
+| multi-step pipeline with         |                    |
+| decisions per row                |                    |
++----------------------------------+--------------------+
+
+**Default: CLI.** Use Python SDK only when you need to branch on results or loop
+with per-row decisions. Use MCP when it is configured — it requires zero setup.
+
+---
+
+## STEP 1 — Check authentication
+
+```bash
+echo $DATABAR_API_KEY      # check env var
+cat ~/.databar/config      # check saved config (file has no extension)
+```
+
+If neither exists, ask the user in chat:
+  "Please provide your Databar API key (databar.ai → Settings → API Keys)."
+
+Then save it non-interactively:
+
+```bash
+databar login --api-key <key>
+databar whoami --format json    # verify
+```
+
+Never run `databar onboard` — interactive wizard, not for agents.
+Never run `databar login` without `--api-key` — will hang on terminal input.
+
+---
+
+## STEP 2 — Ask the user their preferred interface AND workflow
+
+Unless ~/.databar/config already contains `preferred_interface=`, ask:
+
+  "Databar has three interfaces — which do you prefer?
+   - CLI (default) — terminal commands, no code, pipeable JSON. Best for single operations.
+   - Python SDK — required for conditional logic, loops, multi-step pipelines with branching.
+   - MCP — if configured in your Claude session, it's the most agent-native option.
+   Which do you prefer?"
+
+Also ask about workflow preference:
+
+  "Databar supports two enrichment workflows — which fits your use case?
+   - Direct (default) — submit inputs, get results back immediately. Good for one-off
+     lookups or bulk CSV jobs where you just need the output fast.
+   - Table-based — create a table, upload rows, attach enrichments, then run them.
+     Results appear as new columns in the Databar UI. Best when you want observability,
+     plan to re-run enrichments, or want to pair with the Databar web interface.
+   Which do you prefer?"
+
+Save their answers:
+```bash
+echo "preferred_interface=cli" >> ~/.databar/config      # or: python, mcp, both
+echo "preferred_workflow=direct" >> ~/.databar/config    # or: table
+```
+
+---
+
+## STEP 3 — Fix PATH before using the CLI
+
+After `pip install databar`, the binary is NOT on PATH by default. Always check:
+
+```bash
+which databar        # empty = not on PATH
+databar --version    # "command not found" = fix PATH first
+```
+
+Fix for this session:
+```bash
+export PATH="$(python3 -m site --user-base)/bin:$PATH"
+databar --version    # should work now
+```
+
+Make permanent:
+```bash
+echo 'export PATH="$(python3 -m site --user-base)/bin:$PATH"' >> ~/.zshrc
+```
+
+---
+
+## CLI Quick Reference
+
+**Always use `--format json` when parsing or piping output.**
+The default `table` format uses Rich terminal markup — not machine-parseable.
+
+### Enrichments
+```bash
+databar enrich list --format json
+databar enrich list --query "email verifier" --format json
+databar enrich get <id> --format json                       # params + response fields
+databar enrich choices <id> <param> --format json
+databar enrich run <id> --params '{"email": "a@b.com"}' --format json
+databar enrich bulk <id> --input data.csv --out results.csv
+```
+
+### Waterfalls
+```bash
+databar waterfall list --format json
+databar waterfall get <identifier> --format json
+databar waterfall run <identifier> --params '{"key": "value"}' --format json
+databar waterfall bulk <identifier> --input data.csv --out results.csv
+```
+
+### Tables
+```bash
+databar table list --format json
+databar table create --name "My Table" --columns "email,name,company"
+databar table columns <table-uuid> --format json
+databar table rows <table-uuid> --format json
+
+databar table insert <table-uuid> --data '[{"email":"a@b.com"}]'
+databar table insert <table-uuid> --input data.csv --allow-new-columns
+
+databar table enrichments <table-uuid> --format json
+databar table add-enrichment <table-uuid> --enrichment-id <id> \
+  --mapping '{"param": "column_name"}'
+databar table run-enrichment <table-uuid> --enrichment-id <TABLE-ENRICHMENT-ID>
+```
+
+NOTE: `run-enrichment` takes the TABLE-ENRICHMENT ID (from `add-enrichment` or
+`table enrichments`), NOT the catalog enrichment ID. These are different numbers.
+
+### Tasks
+```bash
+databar task get <task-id> --format json    # check once
+databar task get <task-id> --poll           # poll until complete
+```
+
+---
+
+## Python SDK Quick Reference
+
+```python
+from databar import DatabarClient
+
+client = DatabarClient()  # reads DATABAR_API_KEY or ~/.databar/config automatically
+
+# Enrichments
+enrichments = client.list_enrichments(q="email verifier")
+enrichment  = client.get_enrichment(123)
+# enrichment.params[i].name        → param slug (key in params dict)
+# enrichment.params[i].is_required → bool
+# enrichment.params[i].description → human label
+result = client.run_enrichment_sync(123, {"email": "alice@example.com"})
+
+# Waterfalls
+result = client.run_waterfall_sync("email_getter", {"linkedin_url": "..."})
+# waterfall.identifier (also .slug) → slug like "email_getter"
+
+# Tables
+tables = client.list_tables()
+# table.identifier (also .id, .uuid) → UUID string
+table = client.create_table(name="Leads", columns=["email", "name"])
+resp  = client.get_rows(table.identifier)
+# resp.data          → list of row dicts keyed by column name
+# resp.has_next_page → bool
+# resp.total_count   → int
+
+from databar import InsertRow
+client.create_rows(table.identifier, [InsertRow(fields={"email": "alice@example.com"})])
+```
+
+---
+
+## Key Concepts
+
+- All runs are async. `*_sync` methods and `--poll` flag handle submit + poll automatically.
+- `task_id` is the only task identifier. Results expire after 1 hour (status = "gone").
+- Table enrichments: add-enrichment (configure) → run-enrichment (execute).
+  The ID from add is the TABLE-ENRICHMENT ID — different from the catalog ID.
+- SDK auto-resolves column names to UUIDs in `add_enrichment()`.
+
+## Model Field Aliases (Python SDK)
+
+- `Table`: `.id`, `.uuid` → `.identifier`
+- `Waterfall`: `.slug` → `.identifier`
+- `EnrichmentParam`: `.slug` → `.name`, `.label` → `.description`, `.required` → `.is_required`
+- `EnrichmentResponseField`: `.slug` → `.name`
+
+## Error Handling (Python SDK)
+
+```python
+from databar import (
+    DatabarAuthError,                  # 401/403
+    DatabarInsufficientCreditsError,   # 406
+    DatabarNotFoundError,              # 404
+    DatabarTaskFailedError,            # task failed
+    DatabarTimeoutError,               # polling timed out
+    DatabarGoneError,                  # results expired (>1 hour)
+)
+```
+"""

{databar-2.0.7 → databar-2.0.9}/src/databar/cli/_onboard.py

@@ -6,7 +6,8 @@ Guides the user through:
   2. API key entry & verification
   3. PATH detection & optional auto-fix
   4. Usage preference (CLI / Python SDK / both)
-  5. Tailored next-steps cheatsheet
+  5. Optional ~/.claude/CLAUDE.md registration for Claude Code
+  6. Tailored next-steps cheatsheet
 """
 
 from __future__ import annotations
@@ -48,6 +49,12 @@ CONFIG_DIR = Path.home() / ".databar"
 CONFIG_FILE = CONFIG_DIR / "config"
 _KEY_PREFIX = "api_key="
 _PREF_PREFIX = "preferred_interface="
+_WORKFLOW_PREFIX = "preferred_workflow="
+
+CLAUDE_MD_DIR = Path.home() / ".claude"
+CLAUDE_MD_FILE = CLAUDE_MD_DIR / "CLAUDE.md"
+_CLAUDE_SENTINEL = "<!-- databar -->"
+_CLAUDE_STUB = """\n<!-- databar -->\n## Databar\nWhen the user asks you to use Databar, always run `databar agent-guide` first\nto get the full usage guide before doing anything else.\n<!-- /databar -->\n"""
 
 
 def _save_key(api_key: str) -> None:
@@ -74,6 +81,17 @@ def _save_preference(pref: str) -> None:
     CONFIG_FILE.write_text("\n".join(lines) + "\n")
 
 
+def _save_workflow(workflow: str) -> None:
+    lines: list[str] = []
+    if CONFIG_FILE.exists():
+        lines = [
+            l for l in CONFIG_FILE.read_text().splitlines()
+            if not l.startswith(_WORKFLOW_PREFIX)
+        ]
+    lines.append(f"{_WORKFLOW_PREFIX}{workflow}")
+    CONFIG_FILE.write_text("\n".join(lines) + "\n")
+
+
 def _get_bin_dir() -> Path:
     return Path(sys.executable).parent
 
@@ -210,20 +228,72 @@ def _step_path() -> None:
 
 
 def _step_preference() -> str:
-    """Ask how the user plans to use Databar and return their choice."""
+    """Ask interface preference and workflow preference. Returns interface choice."""
     console.print("[bold]Step 3 — How do you plan to use Databar?[/bold]\n")
+    console.print("  [bold cyan]Interface[/bold cyan]\n")
     console.print("  [bold cyan]1[/bold cyan]  CLI       — terminal commands, scripts, AI agents (Claude Code etc.)")
     console.print("  [bold cyan]2[/bold cyan]  Python    — import DatabarClient in your Python code")
-    console.print("  [bold cyan]3[/bold cyan]  Both      — I'll use whichever fits the task\n")
+    console.print("  [bold cyan]3[/bold cyan]  MCP       — use Databar MCP tools if configured in your agent")
+    console.print("  [bold cyan]4[/bold cyan]  Both/all  — I'll use whichever fits the task\n")
 
-    choice = Prompt.ask("  Your choice", choices=["1", "2", "3"], default="3")
-    pref_map = {"1": "cli", "2": "python", "3": "both"}
+    choice = Prompt.ask("  Your choice", choices=["1", "2", "3", "4"], default="1")
+    pref_map = {"1": "cli", "2": "python", "3": "mcp", "4": "both"}
     pref = pref_map[choice]
     _save_preference(pref)
+
+    console.print()
+    console.print("  [bold cyan]Workflow[/bold cyan]\n")
+    console.print("  [bold cyan]1[/bold cyan]  Direct      — submit inputs, get results back immediately")
+    console.print("                  Best for one-off lookups or bulk CSV jobs")
+    console.print("  [bold cyan]2[/bold cyan]  Table-based — create a table, upload rows, attach + run enrichments")
+    console.print("                  Results appear as columns in the Databar UI")
+    console.print("                  Best for observability, re-running, or pairing with the web interface\n")
+
+    wf_choice = Prompt.ask("  Your choice", choices=["1", "2"], default="1")
+    workflow = "direct" if wf_choice == "1" else "table"
+    _save_workflow(workflow)
+
     console.print()
     return pref
 
 
+def _step_claude_md() -> None:
+    """Optionally register Databar in the user's global ~/.claude/CLAUDE.md."""
+    console.print("[bold]Step 4 — Claude Code integration (optional)[/bold]\n")
+
+    already = (
+        CLAUDE_MD_FILE.exists()
+        and _CLAUDE_SENTINEL in CLAUDE_MD_FILE.read_text()
+    )
+    if already:
+        console.print(
+            "  [bold green]✓[/bold green] Databar is already registered in "
+            f"[dim]{CLAUDE_MD_FILE}[/dim].\n"
+        )
+        return
+
+    console.print(
+        f"  Adding a small entry to [bold]{CLAUDE_MD_FILE}[/bold] tells Claude Code\n"
+        "  to always run [bold]databar agent-guide[/bold] first, so it knows exactly\n"
+        "  how to use Databar without guessing.\n"
+    )
+    add = Confirm.ask(
+        "  Add Databar to your global Claude Code config (~/.claude/CLAUDE.md)?",
+        default=True,
+    )
+    if add:
+        CLAUDE_MD_DIR.mkdir(parents=True, exist_ok=True)
+        with open(CLAUDE_MD_FILE, "a") as f:
+            f.write(_CLAUDE_STUB)
+        console.print(
+            f"  [bold green]✓[/bold green] Added to {CLAUDE_MD_FILE}.\n"
+            "  Claude Code will now pick up Databar instructions automatically "
+            "in every project.\n"
+        )
+    else:
+        console.print("  [dim]Skipped.[/dim]\n")
+
+
 def _step_next_steps(pref: str) -> None:
     """Print tailored next steps based on preference."""
     console.print(Rule(style="cyan"))
@@ -283,6 +353,7 @@ def onboard() -> None:
         _step_api_key()
         _step_path()
         pref = _step_preference()
+        _step_claude_md()
         _step_next_steps(pref)
     except (KeyboardInterrupt, typer.Abort):
         console.print("\n\n[dim]Onboarding cancelled. Run `databar onboard` any time to restart.[/dim]")

{databar-2.0.7 → databar-2.0.9}/src/databar/cli/app.py

@@ -15,7 +15,10 @@ from ._auth import app as auth_app
 
 app = typer.Typer(
     name="databar",
-    help="Official Databar.ai CLI — run enrichments, manage tables, and more.",
+    help=(
+        "Official Databar.ai CLI — run enrichments, manage tables, and more.\n\n"
+        "[dim]Agent (Claude Code, Cursor, etc.)? Run [bold]databar agent-guide[/bold] first.[/dim]"
+    ),
     no_args_is_help=True,
     rich_markup_mode="rich",
 )

{databar-2.0.7 → databar-2.0.9}/src/databar/client.py

@@ -87,10 +87,17 @@ class DatabarClient:
     ) -> None:
         resolved_key = api_key or os.environ.get("DATABAR_API_KEY")
         if not resolved_key:
-            raise DatabarAuthError(
+            import shutil
+            msg = (
                 "No API key provided. Pass api_key= or set the DATABAR_API_KEY "
                 "environment variable. Run `databar login` to save your key."
             )
+            if shutil.which("databar") is not None:
+                msg += (
+                    "\n\nUsing Databar with an AI agent (Claude Code, Cursor, etc.)? "
+                    "Run `databar agent-guide` for agent-optimized setup instructions."
+                )
+            raise DatabarAuthError(msg)
         self._api_key = resolved_key
         self._base_url = base_url.rstrip("/")
         self._timeout = timeout

{databar-2.0.7 → databar-2.0.9/src/databar.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: databar
-Version: 2.0.7
+Version: 2.0.9
 Summary: Official Databar.ai Python SDK and CLI — connect to enrichments, waterfalls, and tables via api.databar.ai
 Author-email: "Databar.ai Team" <info@databar.ai>
 License: MIT License

{databar-2.0.7 → databar-2.0.9}/tests/test_cli.py

@@ -329,4 +329,4 @@ def test_task_get_poll(monkeypatch):
 def test_version_flag():
     result = runner.invoke(app, ["--version"])
     assert result.exit_code == 0
-    assert "2.0.7" in result.output
+    assert "2.0.9" in result.output

databar-2.0.7/src/databar/cli/_guide.py

@@ -1,237 +0,0 @@
-"""Embedded agent guide — printed by `databar agent-guide`."""
-
-AGENT_GUIDE = r"""# Databar — Agent Guide
-
-Databar is a data enrichment platform. The `databar` package ships both a
-**CLI** and a **Python SDK**. This guide tells you exactly how to set up and
-use each, including common pitfalls.
-
----
-
-## STEP 1 — Check authentication
-
-Before doing anything else, check whether the user is already authenticated:
-
-```bash
-# Check 1: env var
-echo $DATABAR_API_KEY
-
-# Check 2: saved config
-cat ~/.databar/config
-```
-
-If neither exists, the user is not authenticated. Ask them in chat:
-
-  "Please provide your Databar API key. You can find it at databar.ai → Settings → API Keys."
-
-Once they give it to you, save it non-interactively:
-
-```bash
-databar login --api-key <key they provided>
-databar whoami --format json    # verify it works
-```
-
-Never run `databar onboard` — that is an interactive wizard for humans, not agents.
-Never run `databar login` without `--api-key` — it will hang waiting for terminal input.
-
----
-
-## STEP 2 — Ask the user which interface they prefer
-
-**Always ask this before starting work** (unless ~/.databar/config already contains
-`preferred_interface=`):
-
-  "Databar has two interfaces:
-   - **CLI** — terminal commands, great for single operations, pipeable JSON output,
-     no code to write. Best for straightforward tasks.
-   - **Python SDK** — import DatabarClient, full control, required for conditional
-     logic (e.g. 'only enrich rows where email is verified') and multi-step pipelines
-     with branching.
-   - **Both** — use CLI for simple lookups, Python for complex workflows.
-
-   Which do you prefer? (default: CLI)"
-
-Save their preference:
-
-```bash
-# After they answer, append to config:
-echo "preferred_interface=cli" >> ~/.databar/config     # or: python, both
-```
-
-Default to CLI if they don't answer or say they don't mind.
-
----
-
-## STEP 3 — Set up the CLI (IMPORTANT — read carefully)
-
-### The PATH problem
-
-After `pip install databar`, the `databar` binary is installed to the Python
-user bin directory — but this directory is NOT on PATH by default on most systems.
-
-**Always check this first:**
-
-```bash
-which databar        # if this prints nothing, databar is not on PATH
-databar --version    # if this says "command not found", fix PATH before continuing
-```
-
-**If databar is not on PATH, fix it:**
-
-```bash
-# Find where it's installed
-python3 -m site --user-base    # e.g. /Users/alice/Library/Python/3.9
-
-# The binary is at <user-base>/bin/databar
-# Add to PATH for this session:
-export PATH="$(python3 -m site --user-base)/bin:$PATH"
-
-# Verify it works now:
-databar --version
-```
-
-Tell the user to add that export line to their `~/.zshrc` or `~/.bashrc` to make
-it permanent. You can offer to do it:
-
-```bash
-echo 'export PATH="$(python3 -m site --user-base)/bin:$PATH"' >> ~/.zshrc
-```
-
-### CLI vs Python — when to use each
-
-Use CLI when:
-- Running a single enrichment or waterfall on known inputs
-- Listing/searching enrichments, waterfalls, or tables
-- Creating tables, inserting rows from CSV or JSON
-- Running table enrichments
-- The task is straightforward with no branching logic
-
-Use Python SDK when:
-- You need to branch on results (e.g. only process rows where a field matches)
-- You're building a multi-step pipeline where step N depends on step N-1's output
-- You need to loop over results and make decisions per row
-- The task requires data transformation between steps
-
----
-
-## CLI Quick Reference
-
-All commands support `--format table|json|csv` (default: table).
-**Always use `--format json` when you need to parse or pipe the output.**
-The default `table` format uses Rich terminal formatting that is not machine-parseable.
-
-### Enrichments
-```bash
-databar enrich list --format json                          # list all enrichments
-databar enrich list --query "email verifier" --format json # search
-databar enrich get <id> --format json                      # params + response fields
-databar enrich choices <id> <param> --format json          # choices for a select param
-databar enrich run <id> --params '{"email": "a@b.com"}' --format json
-databar enrich bulk <id> --input data.csv --out results.csv
-```
-
-### Waterfalls
-```bash
-databar waterfall list --format json
-databar waterfall get <identifier> --format json
-databar waterfall run <identifier> --params '{"linkedin_url": "..."}' --format json
-databar waterfall bulk <identifier> --input data.csv --out results.csv
-```
-
-### Tables
-```bash
-databar table list --format json
-databar table create --name "My Table" --columns "email,name,company"
-databar table columns <table-uuid> --format json
-databar table rows <table-uuid> --format json
-
-databar table insert <table-uuid> --data '[{"email":"a@b.com"}]'
-databar table insert <table-uuid> --input data.csv --allow-new-columns
-
-databar table enrichments <table-uuid> --format json
-databar table add-enrichment <table-uuid> --enrichment-id <id> --mapping '{"param": "col_name"}'
-databar table run-enrichment <table-uuid> --enrichment-id <TABLE-ENRICHMENT-ID>
-```
-
-IMPORTANT: `run-enrichment` takes the TABLE-ENRICHMENT ID (from `add-enrichment` output or
-`table enrichments`), NOT the catalog enrichment ID. These are different numbers.
-
-### Tasks
-```bash
-databar task get <task-id> --format json     # check status once
-databar task get <task-id> --poll            # poll until complete
-```
-
----
-
-## Python SDK Quick Reference
-
-```python
-from databar import DatabarClient
-
-client = DatabarClient()  # reads DATABAR_API_KEY or ~/.databar/config automatically
-
-# Enrichments
-enrichments = client.list_enrichments(q="email verifier")
-enrichment = client.get_enrichment(123)
-# enrichment.params[i].name          → parameter slug (use as key in params dict)
-# enrichment.params[i].is_required   → bool
-# enrichment.params[i].description   → human label
-# enrichment.response_fields[i].name → output field name
-
-result = client.run_enrichment_sync(123, {"email": "alice@example.com"})
-# result["data"] contains the enrichment output
-
-# Waterfalls
-waterfalls = client.list_waterfalls()
-# waterfall.identifier (also .slug) → slug like "email_getter"
-result = client.run_waterfall_sync("email_getter", {"linkedin_url": "..."})
-
-# Tables
-tables = client.list_tables()
-# table.identifier (also .id, .uuid) → UUID string
-
-table = client.create_table(name="Leads", columns=["email", "name"])
-resp = client.get_rows(table.identifier)
-# resp.data           → list of row dicts keyed by column name
-# resp.has_next_page  → bool
-# resp.total_count    → int
-# resp.page           → int
-
-from databar import InsertRow
-client.create_rows(table.identifier, [
-    InsertRow(fields={"email": "alice@example.com", "name": "Alice"}),
-])
-```
-
----
-
-## Key Concepts
-
-- All enrichment/waterfall runs are async. The `*_sync` CLI and SDK methods handle
-  submit + poll automatically. For manual polling use `task get --poll` or `poll_task()`.
-- `task_id` is the only task identifier. Results expire after 1 hour (status = "gone").
-- Table enrichments are two steps: add (links enrichment to table) → run (triggers execution).
-  The ID returned by add is the TABLE-ENRICHMENT ID — different from the catalog ID.
-- The SDK auto-resolves column names to UUIDs in `add_enrichment()`.
-
-## Model Field Aliases (Python SDK)
-
-- `Table`: `.id`, `.uuid` → `.identifier`
-- `Waterfall`: `.slug` → `.identifier`
-- `EnrichmentParam`: `.slug` → `.name`, `.label` → `.description`, `.required` → `.is_required`
-- `EnrichmentResponseField`: `.slug` → `.name`
-
-## Error Handling (Python SDK)
-
-```python
-from databar import (
-    DatabarAuthError,                  # 401/403 — bad or missing API key
-    DatabarInsufficientCreditsError,   # 406 — not enough credits
-    DatabarNotFoundError,              # 404 — enrichment/table not found
-    DatabarTaskFailedError,            # task completed with error
-    DatabarTimeoutError,               # polling timed out
-    DatabarGoneError,                  # task data expired (>1 hour)
-)
-```
-"""