parsimony-eia 0.0.1__py3-none-any.whl

parsimony_eia/__init__.py

@@ -0,0 +1,15 @@
+"""US Energy Information Administration (EIA): fetch + catalog enumeration.
+
+API docs: https://www.eia.gov/opendata/documentation.php
+
+The EIA v2 Open Data API is keyed. Supply ``EIA_API_KEY`` (or bind it via
+``load(api_key=...)`` / ``Connector.bind``); a missing key fast-fails with
+:class:`~parsimony.errors.UnauthorizedError`. Register a free key at
+https://www.eia.gov/opendata/register.php.
+"""
+
+from __future__ import annotations
+
+from parsimony_eia.connectors import CONNECTORS, load
+
+__all__ = ["CONNECTORS", "load"]

parsimony_eia/_http.py

@@ -0,0 +1,96 @@
+"""US EIA v2 Open Data API transport: key resolution, client, typed-error chokepoint.
+
+The EIA v2 API is a single keyed REST host. The key rides as an ``?api_key=``
+query param (``api_key`` is in parsimony-core's sensitive-param set, so it is
+auto-redacted from logs) and is also stripped from provenance via
+``secrets=("api_key",)`` on every verb.
+
+``eia_get`` is the per-package mapper chokepoint: EIA returns a clean HTTP 400
+with a useful JSON body for a bad measure / frequency / facet argument
+(``{"error": "Invalid data 'x' ... valid data are 'value'", "code": 400}``).
+``fetch_json`` would collapse that into a generic ``ProviderError(400)`` and drop
+the actionable text, so this chokepoint reads the 400 body and raises
+``InvalidParameterError`` preserving EIA's message; every other status falls
+through to the canonical kernel mappers.
+"""
+
+from __future__ import annotations
+
+import os
+from typing import Any
+
+import httpx
+from parsimony.errors import InvalidParameterError, UnauthorizedError
+from parsimony.transport import HttpClient, map_http_error, map_timeout_error
+from parsimony.transport.helpers import make_http_client
+
+PROVIDER = "eia"
+ENV_VAR = "EIA_API_KEY"
+BASE_URL = "https://api.eia.gov/v2"
+
+# EIA caps every /data page at 5000 rows: a request with ``length`` > 5000 is
+# silently clamped back to 5000 (verified live), so a single call can never
+# return more than 5000 of a dataset's ``total`` rows. The fetch connector pages
+# through with ``offset`` to reach completeness.
+PAGE_SIZE = 5000
+
+# Connectivity for the route-tree enumeration fan-out (~272 nodes). EIA enforces
+# a short-window per-second cap and 429s a 6-wide fan-out; 4 keeps the crawl
+# under that ceiling (the pooled-client retry policy still absorbs the rare 429)
+# so a node is never dropped to a silently-shrunk catalog.
+ENUMERATE_CONCURRENCY = 4
+
+
+def resolve_key(api_key: str) -> str:
+    """Resolve the API key (arg → ``EIA_API_KEY`` env fallback) or fast-fail."""
+    key = api_key or os.environ.get(ENV_VAR, "")
+    if not key:
+        raise UnauthorizedError(PROVIDER, env_var=ENV_VAR)
+    return key
+
+
+def make_eia_client(api_key: str, *, timeout: float = 60.0) -> HttpClient:
+    """Build the EIA client with the key fixed as a (redacted) default query param."""
+    key = resolve_key(api_key)
+    return make_http_client(BASE_URL, query_params={"api_key": key}, timeout=timeout)
+
+
+def _extract_400_message(response: httpx.Response) -> str:
+    """Pull EIA's human-readable error string out of a 400 body (bounded length)."""
+    try:
+        err = response.json().get("error")
+    except ValueError:
+        err = None
+    if isinstance(err, str) and err.strip():
+        return err.strip()[:300]
+    return "invalid request parameter"
+
+
+def eia_get(
+    http: HttpClient,
+    path: str,
+    *,
+    params: dict[str, Any] | None = None,
+    op_name: str,
+) -> dict[str, Any]:
+    """GET ``{BASE_URL}/{path}`` and return the ``response`` object (dict).
+
+    Drops ``None`` params, raises for status, maps a 400 to a message-preserving
+    ``InvalidParameterError`` and every other error family to the canonical typed
+    errors. ``HttpClient.request`` never calls ``raise_for_status`` itself, so we
+    do it here and feed both error families to the kernel mappers.
+    """
+    filtered = {k: v for k, v in (params or {}).items() if v is not None}
+    try:
+        response = http.request("GET", f"/{path.lstrip('/')}", params=filtered or None)
+        response.raise_for_status()
+    except httpx.HTTPStatusError as exc:
+        if exc.response.status_code == 400:
+            raise InvalidParameterError(PROVIDER, _extract_400_message(exc.response)) from exc
+        map_http_error(exc, provider=PROVIDER, op_name=op_name)
+    except httpx.TimeoutException as exc:
+        map_timeout_error(exc, provider=PROVIDER, op_name=op_name)
+
+    body = response.json()
+    inner = body.get("response") if isinstance(body, dict) else None
+    return inner if isinstance(inner, dict) else {}

parsimony_eia/catalog_build.py

@@ -0,0 +1,37 @@
+"""Build the EIA dataset catalog snapshot.
+
+Maintainer tooling, not part of the plugin contract: ``enumerate_eia`` walks the
+v2 route tree to one row per leaf dataset (with its measure/facet manifest), the
+rows become catalog entities, and the catalog is indexed and built. The series
+within a dataset are not catalogued (EIA's ~2M-series universe is the facet
+cartesian product) — they stay fetchable by route+facets or legacy series id.
+"""
+
+from __future__ import annotations
+
+from parsimony.catalog import Catalog
+from parsimony.catalog.policy import discovery_indexes
+from parsimony.catalog.source import entities_from_raw
+
+from parsimony_eia.connectors.enumerate import enumerate_eia
+from parsimony_eia.outputs import EIA_ENUMERATE_OUTPUT
+
+CATALOG_NAMESPACE = "eia"
+
+
+def build_eia_catalog(*, api_key: str | None = None) -> Catalog:
+    """Enumerate every EIA leaf dataset and build a searchable catalog snapshot.
+
+    ``api_key`` falls back to ``EIA_API_KEY`` inside ``enumerate_eia`` (which
+    fast-fails with :class:`~parsimony.errors.UnauthorizedError` if neither is
+    set), so the snapshot can be built straight from the environment.
+    """
+    result = enumerate_eia(api_key=(api_key or "").strip())
+    entries = entities_from_raw(result, EIA_ENUMERATE_OUTPUT)
+    catalog = Catalog(CATALOG_NAMESPACE, indexes=discovery_indexes(entries), default_field="title")
+    catalog.set_entities(entries)
+    catalog.build()
+    return catalog
+
+
+__all__ = ["CATALOG_NAMESPACE", "build_eia_catalog"]

parsimony_eia/connectors/__init__.py

@@ -0,0 +1,19 @@
+"""eia connector registry."""
+
+from __future__ import annotations
+
+from parsimony.connector import Connectors
+
+from parsimony_eia.connectors.enumerate import enumerate_eia
+from parsimony_eia.connectors.fetch import eia_facets, eia_fetch, eia_fetch_series
+from parsimony_eia.search import eia_search
+
+CONNECTORS = Connectors([eia_fetch, eia_fetch_series, eia_facets, enumerate_eia, eia_search])
+
+
+def load(*, api_key: str) -> Connectors:
+    """Return :data:`CONNECTORS` with ``api_key`` bound on every keyed connector."""
+    return CONNECTORS.bind(api_key=api_key)
+
+
+__all__ = ["CONNECTORS", "load"]

parsimony_eia/connectors/enumerate.py

@@ -0,0 +1,209 @@
+"""EIA catalog enumeration: walk the v2 route tree to one row per leaf dataset.
+
+The EIA v2 API is a hierarchy of *route nodes*. A node either lists child
+``routes`` (a category) or terminates as a *leaf dataset* carrying ``data`` (the
+measures), ``facets`` (the dimensions), ``frequency`` and a ``startPeriod`` /
+``endPeriod``. Walking the tree from ``/v2/`` and stopping at the leaves yields
+the authoritative full universe of addressable datasets (archetype B fan-out;
+~232 leaves at the time of writing, ~272 node fetches). Route-node child lists
+are inline and never paginated, so the walk is complete with no list-pagination
+trap.
+
+Each leaf becomes one catalog row whose metadata is a *dimension manifest*: the
+measures it accepts as ``data[0]=`` and the facet ids it accepts as ``facets[]``
+filters, with the measure/facet vocabulary also folded into the indexed
+``description`` so it is BM25-findable. The series within a dataset (the facet
+cartesian product) are not catalogued — they are fetchable by id via
+``eia_fetch`` facet filters, so discovery is at the dataset tier and access is
+total (the BLS/SDMX two-tier split).
+
+Best-effort: a node whose fetch fails is logged and skipped (its subtree is lost
+for that run) so a transient blip yields a partial — not empty — catalog.
+"""
+
+from __future__ import annotations
+
+import logging
+from typing import Any
+
+import httpx
+import pandas as pd
+from parsimony.connector import enumerator
+from parsimony.transport import HttpClient, pooled_client
+
+from parsimony_eia._http import make_eia_client
+from parsimony_eia.outputs import EIA_ENUMERATE_OUTPUT, ENUMERATE_COLUMNS
+
+logger = logging.getLogger(__name__)
+
+# Cap the embedded raw description so a verbose dataset blurb doesn't dominate the
+# indexed text (mirrors parsimony_shared's embedder-friendly cap without adding a
+# dependency for one constant).
+DESCRIPTION_CHAR_CAP = 1500
+
+
+def _get_node(
+    client: HttpClient,
+    route: str,
+) -> dict[str, Any] | None:
+    """Best-effort GET of one route node's metadata. ``None`` on any failure."""
+    path = f"/{route}" if route else "/"
+    try:
+        resp = client.request("GET", path)
+        resp.raise_for_status()
+        body = resp.json()
+    except (httpx.HTTPError, ValueError) as exc:
+        logger.warning("EIA enumerate: node %r failed: %s", route or "<root>", exc)
+        return None
+    inner = body.get("response") if isinstance(body, dict) else None
+    return inner if isinstance(inner, dict) else None
+
+
+def _load_top_routes(client: HttpClient) -> list[str]:
+    """Return the top-level route ids from ``/v2/``.
+
+    This is the **bounding seam** for live tests: monkeypatch it to a 1–2 route
+    slice so the tree walk fires a handful of requests instead of crawling all
+    ~272 nodes. ``enumerate_eia`` calls it by bare name, so the patch takes.
+    """
+    root = _get_node(client, "")
+    if root is None:
+        return []
+    return [str(c["id"]) for c in (root.get("routes") or []) if isinstance(c, dict) and c.get("id")]
+
+
+def _walk(
+    client: HttpClient,
+    routes: list[str],
+    leaves: list[dict[str, Any]],
+) -> None:
+    """Breadth-first serial walk: fetch each node in turn, recurse into children,
+    collect leaf datasets (annotated with their route path) into ``leaves``."""
+    next_level: list[str] = []
+    for route in routes:
+        node = _get_node(client, route)
+        if node is None:
+            continue
+        child_routes = node.get("routes")
+        if isinstance(child_routes, list) and child_routes:
+            next_level.extend(
+                f"{route}/{c['id']}" for c in child_routes if isinstance(c, dict) and c.get("id")
+            )
+        else:
+            node["__route__"] = route
+            leaves.append(node)
+    if next_level:
+        _walk(client, next_level, leaves)
+
+
+def _measure_units(data: dict[str, Any], measures: list[str]) -> list[str]:
+    seen: list[str] = []
+    for m in measures:
+        units = str((data.get(m) or {}).get("units") or "").strip() if isinstance(data.get(m), dict) else ""
+        if units and units not in seen:
+            seen.append(units)
+    return seen
+
+
+def _synthesize_description(
+    *,
+    name: str,
+    category: str,
+    raw_description: str,
+    measures: list[str],
+    units: list[str],
+    facet_pairs: list[tuple[str, str]],
+    frequencies: list[str],
+    start: str,
+    end: str,
+) -> str:
+    """Fold the dataset's query vocabulary into one indexed description string."""
+    parts: list[str] = []
+    if name:
+        parts.append(name)
+    parts.append(f"{category} energy data from the U.S. Energy Information Administration (EIA)")
+    if raw_description:
+        parts.append(raw_description[:DESCRIPTION_CHAR_CAP])
+    if measures:
+        m = ", ".join(measures)
+        parts.append(f"Measures: {m}" + (f" ({', '.join(units)})" if units else ""))
+    if facet_pairs:
+        rendered = ", ".join(f"{fid} ({desc})" if desc else fid for fid, desc in facet_pairs)
+        parts.append(f"Facets: {rendered}")
+    if frequencies:
+        parts.append(f"Frequencies: {', '.join(frequencies)}")
+    if start or end:
+        parts.append(f"Coverage {start or '?'}..{end or '?'}")
+    return ". ".join(p for p in parts if p)
+
+
+def _dataset_row(node: dict[str, Any]) -> dict[str, str]:
+    """Build one catalog row (the dimension manifest) from a leaf dataset node."""
+    route = str(node.get("__route__") or "")
+    name = str(node.get("name") or route)
+    raw_description = str(node.get("description") or "")
+    category = route.split("/", 1)[0]
+
+    data = node.get("data") if isinstance(node.get("data"), dict) else {}
+    measures = [str(k) for k in (data or {})]
+    units = _measure_units(data or {}, measures)
+
+    facets_raw = node.get("facets") if isinstance(node.get("facets"), list) else []
+    facet_pairs: list[tuple[str, str]] = []
+    for f in facets_raw or []:
+        if isinstance(f, dict) and f.get("id"):
+            facet_pairs.append((str(f["id"]), str(f.get("description") or "").strip()))
+    facet_ids = [fid for fid, _ in facet_pairs]
+
+    freqs_raw = node.get("frequency") if isinstance(node.get("frequency"), list) else []
+    frequencies = [str(f["id"]) for f in (freqs_raw or []) if isinstance(f, dict) and f.get("id")]
+
+    start = str(node.get("startPeriod") or "")
+    end = str(node.get("endPeriod") or "")
+
+    return {
+        "code": route,
+        "title": name,
+        "description": _synthesize_description(
+            name=name,
+            category=category,
+            raw_description=raw_description,
+            measures=measures,
+            units=units,
+            facet_pairs=facet_pairs,
+            frequencies=frequencies,
+            start=start,
+            end=end,
+        ),
+        "category": category,
+        "measures": ",".join(measures),
+        "facets": ",".join(facet_ids),
+        "frequencies": ",".join(frequencies),
+        "default_frequency": str(node.get("defaultFrequency") or ""),
+        "start": start,
+        "end": end,
+        "units": ",".join(units),
+    }
+
+
+@enumerator(output=EIA_ENUMERATE_OUTPUT, tags=["macro", "energy", "us"], secrets=("api_key",))
+def enumerate_eia(api_key: str = "") -> pd.DataFrame:
+    """Enumerate every EIA v2 leaf dataset by walking the route tree.
+
+    Emits one row per addressable dataset (KEY=route path) carrying its measure
+    and facet manifest, for catalog indexing. A missing key fast-fails with
+    ``UnauthorizedError`` before any network call.
+    """
+    http = make_eia_client(api_key)
+    leaves: list[dict[str, Any]] = []
+
+    with pooled_client(http) as shared:
+        top_routes = _load_top_routes(shared)
+        if not top_routes:
+            logger.warning("EIA enumerate: /v2/ returned no top-level routes")
+            return pd.DataFrame(columns=list(ENUMERATE_COLUMNS))
+        _walk(shared, top_routes, leaves)
+
+    rows = [_dataset_row(leaf) for leaf in leaves]
+    logger.info("EIA enumerate: %d leaf datasets across %d top-level routes", len(rows), len(top_routes))
+    return pd.DataFrame(rows, columns=list(ENUMERATE_COLUMNS))

parsimony_eia/connectors/fetch.py

@@ -0,0 +1,321 @@
+"""EIA data-fetch connectors.
+
+Three verbs over the EIA v2 API:
+
+* ``eia_fetch`` — fetch a dataset by its route path (``petroleum/pri/spt``) with
+  an optional measure, facet filters and a date window.
+* ``eia_fetch_series`` — fetch by a legacy APIv1 series id (``PET.RWTC.D``,
+  ``ELEC.SALES.CO-RES.A``) via the ``/v2/seriesid/{id}`` path. This addressing
+  scheme lives **outside** the route tree the catalog enumerates, so it is the
+  only connector that reaches a famous series straight from its well-known id.
+* ``eia_facets`` — list the valid values of one facet dimension, so a fetch can
+  be narrowed to a specific series (essential on huge datasets).
+
+**Pagination is mandatory, not optional.** EIA caps every ``/data`` (and
+``/seriesid``) response at 5,000 rows; a dataset like ``petroleum/pri/spt`` daily
+has 91,285 rows and ``electricity/rto/region-data`` hourly has ~18.7M. The
+single-page predecessor silently returned the first 5,000 of whatever matched.
+Both fetch verbs read ``response.total`` and page through with ``offset`` until
+they have it all — guarded by a pre-pagination row-count ceiling that raises an
+actionable ``InvalidParameterError`` (echoing EIA's own "constrain with facet,
+start, end" guidance) so an unbounded request fails loud instead of truncating.
+"""
+
+from __future__ import annotations
+
+import re
+from typing import Annotated, Any
+
+import pandas as pd
+from parsimony.connector import connector
+from parsimony.errors import EmptyDataError, InvalidParameterError
+from parsimony.transport import HttpClient
+
+from parsimony_eia._http import PAGE_SIZE, eia_get, make_eia_client
+from parsimony_eia.outputs import EIA_FACETS_OUTPUT, EIA_FETCH_OUTPUT, EIA_SERIES_OUTPUT
+
+# A single fetch is bounded to this many rows. Above it, EIA's universe is too
+# large to pull whole (electricity hourly alone is ~18.7M rows); we refuse with
+# an actionable narrowing message rather than either truncate silently (the bug
+# we are fixing) or page through millions of rows. ~60 pages.
+MAX_FETCH_ROWS = 300_000
+
+
+def _to_int(value: Any) -> int:
+    try:
+        return int(str(value))
+    except (TypeError, ValueError):
+        return 0
+
+
+def _facet_params(facets: dict[str, Any] | None) -> dict[str, Any]:
+    """Render a ``{facet_id: value | [values]}`` dict into EIA query params.
+
+    ``{"duoarea": "NUS", "product": ["EPCBRENT", "EPD2DC"]}`` →
+    ``facets[duoarea][]=NUS`` and a repeated ``facets[product][]=...``. httpx
+    expands a list-valued param into repeated keys.
+    """
+    out: dict[str, Any] = {}
+    for raw_id, val in (facets or {}).items():
+        fid = str(raw_id).strip()
+        if not fid:
+            continue
+        key = f"facets[{fid}][]"
+        if isinstance(val, (list, tuple, set)):
+            out[key] = [str(v) for v in val]
+        else:
+            out[key] = str(val)
+    return out
+
+
+def _normalize_periods(raw: pd.Series) -> pd.Series:
+    """Coerce EIA period strings (any frequency) to datetimes.
+
+    EIA period shapes by frequency: ``2024`` (annual), ``2024-03`` (monthly),
+    ``2024-03-15`` (daily/weekly), ``2025-Q4`` (quarterly), ``2026-06-10T07``
+    (hourly), ``2026-06-10T03-04`` (local-hourly, trailing TZ band). pandas can
+    parse all but the quarter form, which we expand to its quarter-start month,
+    and the local-hourly TZ band, which we trim to the bare hour.
+    """
+    s = raw.astype("string").str.strip()
+    s = s.str.replace(r"^(\d{4})-Q1$", r"\1-01-01", regex=True)
+    s = s.str.replace(r"^(\d{4})-Q2$", r"\1-04-01", regex=True)
+    s = s.str.replace(r"^(\d{4})-Q3$", r"\1-07-01", regex=True)
+    s = s.str.replace(r"^(\d{4})-Q4$", r"\1-10-01", regex=True)
+    s = s.str.replace(r"(T\d{2})-\d{2}$", r"\1", regex=True)
+    return pd.to_datetime(s, errors="coerce", format="mixed")
+
+
+def _detect_measure_col(df: pd.DataFrame) -> str | None:
+    """Find the measure column when the caller didn't name one (seriesid path).
+
+    EIA names the measure column inconsistently: petroleum uses ``value`` +
+    ``units``; electricity uses ``sales``/``price``/... + a ``<measure>-units``
+    sibling. Prefer ``value``; otherwise the column with a ``{col}-units`` twin.
+    """
+    if "value" in df.columns:
+        return "value"
+    for col in df.columns:
+        if isinstance(col, str) and f"{col}-units" in df.columns:
+            return col
+    return None
+
+
+_DUP_LABEL_RE = re.compile(r"-(name|units|description)$")
+
+
+def _natural_key_columns(df: pd.DataFrame) -> list[str]:
+    """Dimensional columns that identify a row (for boundary-dup-safe dedup).
+
+    EIA offset pagination over an unsorted result is lossless but can repeat a
+    row at a page boundary, so we dedup on the dimensional key: ``period`` plus
+    the facet code columns (dropping the value, the ``*-units``/``*-name``/
+    ``*-description`` label columns, and ``series-description``).
+    """
+    if "series" in df.columns and "period" in df.columns:
+        return ["period", "series"]
+    keys = [
+        c
+        for c in df.columns
+        if isinstance(c, str) and c not in ("value", "units") and not _DUP_LABEL_RE.search(c)
+    ]
+    return keys or list(df.columns)
+
+
+def _shape_observations(
+    data: list[dict[str, Any]],
+    *,
+    key_column: str,
+    key_value: str,
+    title: str,
+    measure: str | None = None,
+) -> pd.DataFrame:
+    """Turn raw EIA rows into the long-format fetch frame.
+
+    Normalizes ``period`` to datetime, the selected/detected measure to a
+    ``value`` numeric column (coercing only that column so string facet metadata
+    survives), dedups boundary-duplicate rows, and stamps the KEY + title. Every
+    other EIA column (facet codes + their ``-name`` labels, ``series``, ``units``)
+    folds in as DATA so a multi-series fetch stays disambiguated.
+    """
+    df = pd.DataFrame(data)
+    if "period" in df.columns:
+        df["period"] = _normalize_periods(df["period"])
+    else:
+        df["period"] = pd.NaT
+
+    measure_col = measure if (measure and measure in df.columns) else _detect_measure_col(df)
+    if measure_col and measure_col != "value":
+        df = df.rename(columns={measure_col: "value"})
+    if "value" in df.columns:
+        df["value"] = pd.to_numeric(df["value"], errors="coerce")
+    else:
+        df["value"] = pd.NA
+
+    df = df.drop_duplicates(subset=_natural_key_columns(df), keep="first").reset_index(drop=True)
+
+    df[key_column] = key_value
+    df["title"] = title
+    return df
+
+
+def _paginate(
+    http: HttpClient,
+    path: str,
+    base_params: dict[str, Any],
+    *,
+    op_name: str,
+) -> tuple[list[dict[str, Any]], dict[str, Any]]:
+    """Page through ALL rows for a /data or /seriesid request.
+
+    Returns ``(rows, first_response)`` — the full row list plus the first page's
+    response envelope (for its ``description``). Raises ``InvalidParameterError``
+    before paging if the match exceeds ``MAX_FETCH_ROWS``.
+    """
+    first = eia_get(
+        http, path, params={**base_params, "offset": 0, "length": PAGE_SIZE}, op_name=op_name
+    )
+    total = _to_int(first.get("total"))
+    rows: list[dict[str, Any]] = list(first.get("data") or [])
+
+    if total > MAX_FETCH_ROWS:
+        raise InvalidParameterError(
+            "eia",
+            f"this request matches {total} rows, above the {MAX_FETCH_ROWS}-row fetch ceiling; "
+            "narrow it with facets=, frequency=, start= or end= "
+            "(EIA caps every page at 5000 rows).",
+        )
+
+    offset = len(rows)
+    while offset < total:
+        page = eia_get(
+            http, path, params={**base_params, "offset": offset, "length": PAGE_SIZE}, op_name=op_name
+        )
+        chunk = list(page.get("data") or [])
+        if not chunk:
+            break
+        rows.extend(chunk)
+        offset += len(chunk)
+
+    return rows, first
+
+
+@connector(output=EIA_FETCH_OUTPUT, tags=["macro", "energy", "us"], secrets=("api_key",))
+def eia_fetch(
+    route: Annotated[str, "ns:eia"],
+    measure: str = "value",
+    facets: dict[str, Any] | None = None,
+    frequency: str | None = None,
+    start: str | None = None,
+    end: str | None = None,
+    api_key: str = "",
+) -> pd.DataFrame:
+    """Fetch an EIA dataset by route path (e.g. ``petroleum/pri/spt``).
+
+    ``measure`` selects EIA's required ``data[0]=`` facet — it is route-specific
+    (``value`` for most series; ``price``/``sales``/``revenue``/``customers`` for
+    electricity, ``heat-content`` etc. for coal). ``facets`` is a
+    ``{facet_id: value | [values]}`` dict (read the dataset's facet ids from the
+    catalog, and their valid values from ``eia_facets``) that narrows the result
+    to a series. The full result is paged in (EIA's 5,000-row page cap is handled
+    internally); a match above the row ceiling raises ``InvalidParameterError``
+    asking you to narrow it.
+    """
+    r = route.strip()
+    if not r:
+        raise InvalidParameterError("eia", "route must be non-empty")
+    m = measure.strip()
+    if not m:
+        raise InvalidParameterError("eia", "measure must be non-empty")
+
+    http = make_eia_client(api_key)
+    base_params: dict[str, Any] = {
+        "data[0]": m,
+        "frequency": frequency,
+        "start": start,
+        "end": end,
+        **_facet_params(facets),
+    }
+    rows, first = _paginate(http, f"{r}/data", base_params, op_name="eia_fetch")
+    if not rows:
+        raise EmptyDataError("eia", query_params={"route": r, "measure": m})
+
+    title = str(first.get("description") or r)
+    return _shape_observations(rows, key_column="route", key_value=r, title=title, measure=m)
+
+
+@connector(output=EIA_SERIES_OUTPUT, tags=["macro", "energy", "us"], secrets=("api_key",))
+def eia_fetch_series(
+    series_id: Annotated[str, "ns:eia"],
+    start: str | None = None,
+    end: str | None = None,
+    api_key: str = "",
+) -> pd.DataFrame:
+    """Fetch an EIA series by its legacy APIv1 series id (e.g. ``PET.RWTC.D``).
+
+    Uses the ``/v2/seriesid/{id}`` compatibility path, which reaches a fully
+    specified series straight from its well-known id — the addressing scheme used
+    across the EIA/FRED ecosystem and EIA's own data browsers. This path lives
+    outside the route tree the catalog indexes, so it is the way to retrieve a
+    famous series (WTI crude ``PET.RWTC.D``, Henry Hub ``NG.RNGWHHD.D``) without
+    composing a route + facet query. Paged in full like ``eia_fetch``.
+    """
+    sid = series_id.strip()
+    if not sid:
+        raise InvalidParameterError("eia", "series_id must be non-empty")
+
+    http = make_eia_client(api_key)
+    base_params: dict[str, Any] = {"start": start, "end": end}
+    rows, first = _paginate(http, f"seriesid/{sid}", base_params, op_name="eia_fetch_series")
+    if not rows:
+        raise EmptyDataError("eia", query_params={"series_id": sid})
+
+    # A single series id resolves to one series, so its per-row series-description
+    # (e.g. "Cushing, OK WTI Spot Price FOB") is a far better title than the
+    # generic dataset-level `description`; fall back to that, then the id.
+    series_desc = next(
+        (str(row["series-description"]) for row in rows if str(row.get("series-description") or "").strip()),
+        "",
+    )
+    title = series_desc or str(first.get("description") or sid)
+    return _shape_observations(rows, key_column="series_id", key_value=sid, title=title)
+
+
+@connector(output=EIA_FACETS_OUTPUT, tags=["macro", "energy", "us"], secrets=("api_key",))
+def eia_facets(
+    route: Annotated[str, "ns:eia"],
+    facet: str,
+    api_key: str = "",
+) -> pd.DataFrame:
+    """List the valid values of one facet dimension of an EIA dataset.
+
+    Given a dataset ``route`` and one of its facet ids (from the catalog's
+    ``facets`` metadata), returns the ``{id, name}`` value vocabulary so a fetch
+    can be narrowed to a specific series. This is the bridge that makes huge
+    datasets usable — e.g. ``electricity/rto/region-data`` is ~18.7M rows, so an
+    agent must narrow by ``respondent``/``fueltype`` facet values, which it
+    discovers here rather than by a blind (ceiling-rejected) full fetch.
+    """
+    r = route.strip()
+    f = facet.strip()
+    if not r:
+        raise InvalidParameterError("eia", "route must be non-empty")
+    if not f:
+        raise InvalidParameterError("eia", "facet must be non-empty")
+
+    http = make_eia_client(api_key)
+    resp = eia_get(http, f"{r}/facet/{f}", op_name="eia_facets")
+    values = resp.get("facets") or []
+    rows = [
+        {
+            "facet_value": str(v.get("id", "")),
+            "name": str(v.get("name") or v.get("id", "")),
+            "facet": f,
+            "route": r,
+        }
+        for v in values
+        if isinstance(v, dict) and v.get("id")
+    ]
+    if not rows:
+        raise EmptyDataError("eia", query_params={"route": r, "facet": f})
+    return pd.DataFrame(rows)

parsimony_eia/outputs.py

@@ -0,0 +1,81 @@
+"""EIA connector output schemas.
+
+The catalog row is one per **leaf dataset** (the route-tree terminal node). Its
+metadata carries the dataset's *query vocabulary* — the measures it accepts as
+``data[0]=`` and the facet dimensions it accepts as ``facets[...]`` filters — so
+an agent can read a search hit and construct a precise ``eia_fetch`` /
+``eia_facets`` call without a blind round-trip (the SDMX/BLS dimension-manifest
+pattern). The ``code`` KEY holds the route path (e.g. ``petroleum/pri/spt``).
+"""
+
+from __future__ import annotations
+
+from parsimony.result import Column, ColumnRole, OutputConfig
+
+EIA_ENUMERATE_OUTPUT = OutputConfig(
+    columns=[
+        Column(name="code", role=ColumnRole.KEY, namespace="eia"),
+        Column(name="title", role=ColumnRole.TITLE),
+        Column(name="description", role=ColumnRole.METADATA),
+        Column(name="category", role=ColumnRole.METADATA),
+        Column(name="measures", role=ColumnRole.METADATA),
+        Column(name="facets", role=ColumnRole.METADATA),
+        Column(name="frequencies", role=ColumnRole.METADATA),
+        Column(name="default_frequency", role=ColumnRole.METADATA),
+        Column(name="start", role=ColumnRole.METADATA),
+        Column(name="end", role=ColumnRole.METADATA),
+        Column(name="units", role=ColumnRole.METADATA),
+    ]
+)
+
+ENUMERATE_COLUMNS: tuple[str, ...] = (
+    "code",
+    "title",
+    "description",
+    "category",
+    "measures",
+    "facets",
+    "frequencies",
+    "default_frequency",
+    "start",
+    "end",
+    "units",
+)
+
+# Fetch returns long-format observations. ``route`` is the KEY (param_key=route
+# links a catalog hit back to the fetch parameter); the selected measure is
+# normalized to ``value``; every other EIA column (facet codes + their ``-name``
+# labels, ``series``, ``series-description``, ``units``) folds in as DATA so a
+# multi-series fetch stays disambiguated.
+EIA_FETCH_OUTPUT = OutputConfig(
+    columns=[
+        Column(name="route", role=ColumnRole.KEY, namespace="eia"),
+        Column(name="title", role=ColumnRole.TITLE),
+        Column(name="period", dtype="datetime", role=ColumnRole.DATA),
+        Column(name="value", dtype="numeric", role=ColumnRole.DATA),
+    ]
+)
+
+# Fetch by legacy APIv1 series id (the out-of-tree `/v2/seriesid/{id}` path —
+# e.g. `PET.RWTC.D`, `ELEC.SALES.CO-RES.A`). Same observation shape as the route
+# fetch; the series id fully specifies the measure, which is normalized to value.
+EIA_SERIES_OUTPUT = OutputConfig(
+    columns=[
+        Column(name="series_id", role=ColumnRole.KEY, namespace="eia"),
+        Column(name="title", role=ColumnRole.TITLE),
+        Column(name="period", dtype="datetime", role=ColumnRole.DATA),
+        Column(name="value", dtype="numeric", role=ColumnRole.DATA),
+    ]
+)
+
+# Facet-value discovery: one row per valid value of a dataset's facet dimension,
+# so an agent can narrow a fetch to a specific series (essential on huge datasets
+# — electricity hourly is ~18M rows). KEY=facet_value, TITLE=name.
+EIA_FACETS_OUTPUT = OutputConfig(
+    columns=[
+        Column(name="facet_value", role=ColumnRole.KEY, namespace="eia"),
+        Column(name="name", role=ColumnRole.TITLE),
+        Column(name="facet", role=ColumnRole.METADATA),
+        Column(name="route", role=ColumnRole.METADATA),
+    ]
+)

parsimony_eia/search.py

@@ -0,0 +1,39 @@
+"""Semantic search over the published EIA dataset catalog."""
+
+from __future__ import annotations
+
+from parsimony.catalog.search import CatalogSearchParams, make_local_search_connector
+from parsimony.result import Column, ColumnRole, OutputConfig
+
+from parsimony_eia.catalog_build import build_eia_catalog
+
+EiaSearchParams = CatalogSearchParams
+
+PARSIMONY_EIA_CATALOG_URL_ENV = "PARSIMONY_EIA_CATALOG_URL"
+
+EIA_SEARCH_OUTPUT = OutputConfig(
+    columns=[
+        Column(name="code", role=ColumnRole.KEY, namespace="eia"),
+        Column(name="title", role=ColumnRole.TITLE),
+        Column(name="score", role=ColumnRole.DATA),
+    ]
+)
+
+eia_search = make_local_search_connector(
+    provider="eia",
+    default_url="hf://parsimony-dev/eia",
+    catalog_url_env_var=PARSIMONY_EIA_CATALOG_URL_ENV,
+    build_catalog=build_eia_catalog,
+    tags=["macro", "energy", "us", "tool"],
+    description=(
+        "Semantic-search the U.S. Energy Information Administration (EIA) Open Data catalog of "
+        "energy datasets (petroleum, natural gas, electricity, coal, nuclear, renewables, "
+        "emissions, international). Each hit is a dataset route plus its measure and facet "
+        "vocabulary; pass the route to eia_fetch(route=..., measure=..., facets=...) for "
+        "observations, eia_facets(route=..., facet=...) to list a facet's values, or use "
+        "eia_fetch_series(series_id=...) for a known legacy series id like PET.RWTC.D."
+    ),
+    output_columns=EIA_SEARCH_OUTPUT.columns,
+)
+
+__all__ = ["PARSIMONY_EIA_CATALOG_URL_ENV", "EiaSearchParams", "eia_search"]

parsimony_eia-0.0.1.dist-info/METADATA

@@ -0,0 +1,112 @@
+Metadata-Version: 2.4
+Name: parsimony-eia
+Version: 0.0.1
+Summary: U.S. Energy Information Administration connector for the parsimony framework
+Project-URL: Homepage, https://www.eia.gov
+Project-URL: Repository, https://github.com/ockham-sh/parsimony-connectors
+Project-URL: Issues, https://github.com/ockham-sh/parsimony-connectors/issues
+Author-email: "Ockham.sh" <team@ockham.sh>
+License-Expression: Apache-2.0
+License-File: LICENSE
+Keywords: connectors,data,eia,finance,parsimony
+Classifier: Development Status :: 4 - Beta
+Classifier: Intended Audience :: Developers
+Classifier: Intended Audience :: Financial and Insurance Industry
+Classifier: License :: OSI Approved :: Apache Software License
+Classifier: Programming Language :: Python :: 3
+Classifier: Programming Language :: Python :: 3.11
+Classifier: Programming Language :: Python :: 3.12
+Classifier: Programming Language :: Python :: 3.13
+Classifier: Topic :: Office/Business :: Financial
+Classifier: Topic :: Software Development :: Libraries :: Python Modules
+Classifier: Typing :: Typed
+Requires-Python: >=3.11
+Requires-Dist: pandas<3,>=2.3.0
+Requires-Dist: parsimony-core>=0.0.1
+Requires-Dist: pydantic<3,>=2.11.1
+Provides-Extra: dev
+Requires-Dist: mypy>=1.10; extra == 'dev'
+Requires-Dist: pytest-cov>=5.0; extra == 'dev'
+Requires-Dist: pytest>=9.0.3; extra == 'dev'
+Requires-Dist: respx>=0.22.0; extra == 'dev'
+Requires-Dist: ruff>=0.15.10; extra == 'dev'
+Description-Content-Type: text/markdown
+
+# parsimony-eia
+
+U.S. Energy Information Administration (EIA) connector — fetches energy data (petroleum, electricity, natural gas, coal, renewables) from the EIA v2 Open Data API.
+
+Part of the [parsimony-connectors](https://github.com/ockham-sh/parsimony-connectors) monorepo. Distributed standalone on PyPI as `parsimony-eia`.
+
+## Connectors
+
+| Name | Kind | Description |
+|---|---|---|
+| `eia_search` | tool | Semantic-search the dataset catalog; returns a dataset `route` to pass to `eia_fetch`. |
+| `eia_fetch` | fetch | Fetch a dataset by `route` (e.g. `petroleum/pri/spt`) with optional `measure`, `facets`, `frequency`, `start`, `end`. Paged in full. |
+| `eia_fetch_series` | fetch | Fetch by a legacy APIv1 series id (e.g. `PET.RWTC.D`) via the `/v2/seriesid/{id}` path. |
+| `eia_facets` | tool | List the valid `{id, name}` values of a dataset's facet dimension, to narrow a fetch to a series. |
+| `enumerate_eia` | enumerator | Walk the v2 route tree to one row per leaf dataset (the catalog feed). |
+
+## Coverage
+
+The catalog indexes **all 232 leaf datasets** across EIA's 14 top-level
+categories (petroleum, natural gas, electricity, coal, nuclear, renewables/
+densified-biomass, total energy, emissions, international, SEDS, STEO, AEO, IEO,
+crude-oil imports), each carrying its measure and facet-dimension manifest. EIA's
+full series universe (~2M series) is the facet cartesian product of those
+datasets — too large to catalog individually, but every series is **fetchable**:
+by `route` + `facets` filters, or by its legacy series id via `eia_fetch_series`.
+So discovery is at the dataset tier and access is total.
+
+EIA caps every data response at 5,000 rows; the fetch connectors page through to
+completeness automatically (a request matching more than 300,000 rows is refused
+with guidance to narrow it). Data is U.S. federal public domain — cite as
+"Source: U.S. Energy Information Administration".
+
+## Install
+
+```bash
+pip install parsimony-eia
+```
+
+Pulls in a compatible `parsimony-core` automatically. Verify discovery:
+
+```bash
+python -c "from parsimony import discover; print([p.name for p in discover.iter_providers()])"
+```
+
+## Configuration
+
+Set the following environment variable:
+
+```bash
+export EIA_API_KEY="<your-key>"
+```
+
+Get a free key at https://www.eia.gov/opendata/register.php.
+
+## Quick start
+
+```python
+from parsimony_eia import CONNECTORS
+
+result = CONNECTORS["eia_fetch"](route="petroleum/pri/spt", frequency="monthly")
+print(result.data.head())
+```
+
+For multi-plugin composition:
+
+```python
+from parsimony import discover
+connectors = discover.load_all()
+```
+
+## Provider
+
+- Homepage: https://www.eia.gov
+- API docs: https://www.eia.gov/opendata/documentation.php
+
+## License
+
+See [LICENSE](./LICENSE).

parsimony_eia-0.0.1.dist-info/RECORD

@@ -0,0 +1,14 @@
+parsimony_eia/__init__.py,sha256=9ASwtwLQ4K69zxYaTMFj02s7OzAr3ZvhqgRC6jXcKNw,525
+parsimony_eia/_http.py,sha256=D0z4fBy1PipVVOcBtuOaSc3Pv_UgYUeX07e6LfUSoDk,3991
+parsimony_eia/catalog_build.py,sha256=VUphJj3j7wLK45clGX5hGO5oyn5fLn21Q0c1fJYxErc,1508
+parsimony_eia/outputs.py,sha256=FsqADK6fkSmwTt9Lq-Cnez0nIr5j7x9-MIXg5tltQqo,3303
+parsimony_eia/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+parsimony_eia/search.py,sha256=vNsl4xl1-P9jApXKc5Au2DlDuvW79jWKkOn7S88t3-I,1579
+parsimony_eia/connectors/__init__.py,sha256=ikVfyurz9wNFSqktx1qZ8MIpQRRXZIfwgg8zfCF_98E,601
+parsimony_eia/connectors/enumerate.py,sha256=lErSaOrfMYc9z0NIlt4-6mydBsaXrj-ZE34navJfgxM,8085
+parsimony_eia/connectors/fetch.py,sha256=jWpRaxXzNvk6Jo3sy-x1k_Z6eCHpUAluz4RQQpfozV4,12772
+parsimony_eia-0.0.1.dist-info/METADATA,sha256=E2qizIje_f83954IxF7QRPEWSJi5pUJjPswRuwZA6lU,4216
+parsimony_eia-0.0.1.dist-info/WHEEL,sha256=mffPy8wBnZQn2VnJUU5jE99KsxaSfiyMHV9Yt0aLVxs,87
+parsimony_eia-0.0.1.dist-info/entry_points.txt,sha256=qJPVlobYtY-spSyALdzvxR-aFGE43NXOPopKOA-Fluw,42
+parsimony_eia-0.0.1.dist-info/licenses/LICENSE,sha256=PtHUFTCSwal_QX2Ijk2cx_bpsPV6ooZUMCYAxKBHNu0,10760
+parsimony_eia-0.0.1.dist-info/RECORD,,

parsimony_eia-0.0.1.dist-info/WHEEL

@@ -0,0 +1,4 @@
+Wheel-Version: 1.0
+Generator: hatchling 1.30.1
+Root-Is-Purelib: true
+Tag: py3-none-any

parsimony_eia-0.0.1.dist-info/entry_points.txt

@@ -0,0 +1,2 @@
+[parsimony.providers]
+eia = parsimony_eia

parsimony_eia-0.0.1.dist-info/licenses/LICENSE

@@ -0,0 +1,190 @@
+                                 Apache License
+                           Version 2.0, January 2004
+                        http://www.apache.org/licenses/
+
+   TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION
+
+   1. Definitions.
+
+      "License" shall mean the terms and conditions for use, reproduction,
+      and distribution as defined by Sections 1 through 9 of this document.
+
+      "Licensor" shall mean the copyright owner or entity authorized by
+      the copyright owner that is granting the License.
+
+      "Legal Entity" shall mean the union of the acting entity and all
+      other entities that control, are controlled by, or are under common
+      control with that entity. For the purposes of this definition,
+      "control" means (i) the power, direct or indirect, to cause the
+      direction or management of such entity, whether by contract or
+      otherwise, or (ii) ownership of fifty percent (50%) or more of the
+      outstanding shares, or (iii) beneficial ownership of such entity.
+
+      "You" (or "Your") shall mean an individual or Legal Entity
+      exercising permissions granted by this License.
+
+      "Source" form shall mean the preferred form for making modifications,
+      including but not limited to software source code, documentation
+      source, and configuration files.
+
+      "Object" form shall mean any form resulting from mechanical
+      transformation or translation of a Source form, including but
+      not limited to compiled object code, generated documentation,
+      and conversions to other media types.
+
+      "Work" shall mean the work of authorship, whether in Source or
+      Object form, made available under the License, as indicated by a
+      copyright notice that is included in or attached to the work
+      (an example is provided in the Appendix below).
+
+      "Derivative Works" shall mean any work, whether in Source or Object
+      form, that is based on (or derived from) the Work and for which the
+      editorial revisions, annotations, elaborations, or other modifications
+      represent, as a whole, an original work of authorship. For the purposes
+      of this License, Derivative Works shall not include works that remain
+      separable from, or merely link (or bind by name) to the interfaces of,
+      the Work and Derivative Works thereof.
+
+      "Contribution" shall mean any work of authorship, including
+      the original version of the Work and any modifications or additions
+      to that Work or Derivative Works thereof, that is intentionally
+      submitted to the Licensor for inclusion in the Work by the copyright owner
+      or by an individual or Legal Entity authorized to submit on behalf of
+      the copyright owner. For the purposes of this definition, "submitted"
+      means any form of electronic, verbal, or written communication sent
+      to the Licensor or its representatives, including but not limited to
+      communication on electronic mailing lists, source code control systems,
+      and issue tracking systems that are managed by, or on behalf of, the
+      Licensor for the purpose of discussing and improving the Work, but
+      excluding communication that is conspicuously marked or otherwise
+      designated in writing by the copyright owner as "Not a Contribution."
+
+      "Contributor" shall mean Licensor and any individual or Legal Entity
+      on behalf of whom a Contribution has been received by the Licensor and
+      subsequently incorporated within the Work.
+
+   2. Grant of Copyright License. Subject to the terms and conditions of
+      this License, each Contributor hereby grants to You a perpetual,
+      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+      copyright license to reproduce, prepare Derivative Works of,
+      publicly display, publicly perform, sublicense, and distribute the
+      Work and such Derivative Works in Source or Object form.
+
+   3. Grant of Patent License. Subject to the terms and conditions of
+      this License, each Contributor hereby grants to You a perpetual,
+      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+      (except as stated in this section) patent license to make, have made,
+      use, offer to sell, sell, import, and otherwise transfer the Work,
+      where such license applies only to those patent claims licensable
+      by such Contributor that are necessarily infringed by their
+      Contribution(s) alone or by combination of their Contribution(s)
+      with the Work to which such Contribution(s) was submitted. If You
+      institute patent litigation against any entity (including a
+      cross-claim or counterclaim in a lawsuit) alleging that the Work
+      or a Contribution incorporated within the Work constitutes direct
+      or contributory patent infringement, then any patent licenses
+      granted to You under this License for that Work shall terminate
+      as of the date such litigation is filed.
+
+   4. Redistribution. You may reproduce and distribute copies of the
+      Work or Derivative Works thereof in any medium, with or without
+      modifications, and in Source or Object form, provided that You
+      meet the following conditions:
+
+      (a) You must give any other recipients of the Work or
+          Derivative Works a copy of this License; and
+
+      (b) You must cause any modified files to carry prominent notices
+          stating that You changed the files; and
+
+      (c) You must retain, in the Source form of any Derivative Works
+          that You distribute, all copyright, patent, trademark, and
+          attribution notices from the Source form of the Work,
+          excluding those notices that do not pertain to any part of
+          the Derivative Works; and
+
+      (d) If the Work includes a "NOTICE" text file as part of its
+          distribution, then any Derivative Works that You distribute must
+          include a readable copy of the attribution notices contained
+          within such NOTICE file, excluding any notices that do not
+          pertain to any part of the Derivative Works, in at least one
+          of the following places: within a NOTICE text file distributed
+          as part of the Derivative Works; within the Source form or
+          documentation, if provided along with the Derivative Works; or,
+          within a display generated by the Derivative Works, if and
+          wherever such third-party notices normally appear. The contents
+          of the NOTICE file are for informational purposes only and
+          do not modify the License. You may add Your own attribution
+          notices within Derivative Works that You distribute, alongside
+          or as an addendum to the NOTICE text from the Work, provided
+          that such additional attribution notices cannot be construed
+          as modifying the License.
+
+      You may add Your own copyright statement to Your modifications and
+      may provide additional or different license terms and conditions
+      for use, reproduction, or distribution of Your modifications, or
+      for any such Derivative Works as a whole, provided Your use,
+      reproduction, and distribution of the Work otherwise complies with
+      the conditions stated in this License.
+
+   5. Submission of Contributions. Unless You explicitly state otherwise,
+      any Contribution intentionally submitted for inclusion in the Work
+      by You to the Licensor shall be under the terms and conditions of
+      this License, without any additional terms or conditions.
+      Notwithstanding the above, nothing herein shall supersede or modify
+      the terms of any separate license agreement you may have executed
+      with Licensor regarding such Contributions.
+
+   6. Trademarks. This License does not grant permission to use the trade
+      names, trademarks, service marks, or product names of the Licensor,
+      except as required for reasonable and customary use in describing the
+      origin of the Work and reproducing the content of the NOTICE file.
+
+   7. Disclaimer of Warranty. Unless required by applicable law or
+      agreed to in writing, Licensor provides the Work (and each
+      Contributor provides its Contributions) on an "AS IS" BASIS,
+      WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
+      implied, including, without limitation, any warranties or conditions
+      of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A
+      PARTICULAR PURPOSE. You are solely responsible for determining the
+      appropriateness of using or redistributing the Work and assume any
+      risks associated with Your exercise of permissions under this License.
+
+   8. Limitation of Liability. In no event and under no legal theory,
+      whether in tort (including negligence), contract, or otherwise,
+      unless required by applicable law (such as deliberate and grossly
+      negligent acts) or agreed to in writing, shall any Contributor be
+      liable to You for damages, including any direct, indirect, special,
+      incidental, or consequential damages of any character arising as a
+      result of this License or out of the use or inability to use the
+      Work (including but not limited to damages for loss of goodwill,
+      work stoppage, computer failure or malfunction, or any and all
+      other commercial damages or losses), even if such Contributor
+      has been advised of the possibility of such damages.
+
+   9. Accepting Warranty or Additional Liability. While redistributing
+      the Work or Derivative Works thereof, You may choose to offer,
+      and charge a fee for, acceptance of support, warranty, indemnity,
+      or other liability obligations and/or rights consistent with this
+      License. However, in accepting such obligations, You may act only
+      on Your own behalf and on Your sole responsibility, not on behalf
+      of any other Contributor, and only if You agree to indemnify,
+      defend, and hold each Contributor harmless for any liability
+      incurred by, or claims asserted against, such Contributor by reason
+      of your accepting any such warranty or additional liability.
+
+   END OF TERMS AND CONDITIONS
+
+   Copyright 2026 Ockham.sh
+
+   Licensed under the Apache License, Version 2.0 (the "License");
+   you may not use this file except in compliance with the License.
+   You may obtain a copy of the License at
+
+       http://www.apache.org/licenses/LICENSE-2.0
+
+   Unless required by applicable law or agreed to in writing, software
+   distributed under the License is distributed on an "AS IS" BASIS,
+   WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+   See the License for the specific language governing permissions and
+   limitations under the License.