"""Canonical Polars schemas for every published config."""

from __future__ import annotations

from collections.abc import Mapping

import polars as pl

UTC_DATETIME = pl.Datetime(time_unit="us", time_zone="UTC")

BUILD_VERSION = "1.0.0"
PARSER_VERSION = "1.0.0"

# One row per public information arrival. These are the kinds.
EVENT_TYPES = (
    "STATEMENT",
    "IMPLEMENTATION_NOTE",
    "SEP",
    "MINUTES",
    "PRESS_CONFERENCE",
    "LONGER_RUN_STRATEGY",
    "OTHER_POLICY_DOCUMENT",
)
DOCUMENT_TYPES = (
    "statement",
    "implementation_note",
    "minutes",
    "sep_text",
    "press_conference_transcript",
    "longer_run_strategy",
    "other",
)
MEETING_TYPES = ("scheduled", "unscheduled", "notation_vote", "conference_call", "other")
# How precisely the moment of publication is known. `exact` means the document
# stated it; `date` means only the day is known and the instant is the end of
# that day in Washington, which is late and never early.
KNOWLEDGE_PRECISIONS = ("exact", "minute", "hour", "date", "modeled_conservative", "unknown")
STATUSES = (
    "scheduled", "discovered", "fetched", "parsed", "validated", "published",
    "quarantined", "revised", "cancelled",
)
DECISIONS = ("raise", "cut", "hold", "other")
VOTES = ("for", "against", "not_voting", "unknown")

PROVENANCE: dict[str, pl.DataType] = {
    "source": pl.String,
    "source_url": pl.String,
    "source_document_id": pl.String,
    "source_type": pl.String,
    "raw_sha256": pl.String,
    "parser_version": pl.String,
    "first_seen_at": UTC_DATETIME,
    "ingested_at": UTC_DATETIME,
}

MEETING_SCHEMA: dict[str, pl.DataType] = {
    "meeting_id": pl.String,
    "meeting_start_date": pl.Date,
    "meeting_end_date": pl.Date,
    "meeting_type": pl.String,
    "scheduled": pl.Boolean,
    "has_statement": pl.Boolean,
    "has_implementation_note": pl.Boolean,
    "has_minutes": pl.Boolean,
    "has_sep": pl.Boolean,
    "has_press_conference": pl.Boolean,
    "chair": pl.String,
    # The instant the first thing about this meeting became public. A meetings
    # table without one cannot be mounted, and indexing it on the meeting date
    # would be exactly the look-ahead this dataset exists to remove.
    "knowledge_date": UTC_DATETIME,
    "source_url": pl.String,
    "ingested_at": UTC_DATETIME,
}

EVENT_SCHEMA: dict[str, pl.DataType] = {
    "event_id": pl.String,
    "meeting_id": pl.String,
    "event_type": pl.String,
    "scheduled_at": UTC_DATETIME,
    "release_at": UTC_DATETIME,
    "knowledge_at": UTC_DATETIME,
    # The same instant under the name ziplime indexes on.
    "knowledge_date": UTC_DATETIME,
    "release_date_et": pl.Date,
    "release_time_et": pl.Time,
    "timezone": pl.String,
    "target_rate_lower": pl.Float64,
    "target_rate_upper": pl.Float64,
    "target_rate_mid": pl.Float64,
    "rate_change_bp": pl.Float64,
    # The administered rates, from the implementation note published with the
    # statement. They are levels rather than a range, and the front end trades
    # against them directly.
    "iorb_rate": pl.Float64,
    "on_rrp_rate": pl.Float64,
    "primary_credit_rate": pl.Float64,
    "decision": pl.String,
    "vote_for": pl.Int32,
    "vote_against": pl.Int32,
    "knowledge_time_precision": pl.String,
    "status": pl.String,
    **PROVENANCE,
}

DOCUMENT_SCHEMA: dict[str, pl.DataType] = {
    "document_id": pl.String,
    "meeting_id": pl.String,
    "event_id": pl.String,
    "document_type": pl.String,
    "release_at": UTC_DATETIME,
    "knowledge_at": UTC_DATETIME,
    "knowledge_date": UTC_DATETIME,
    "knowledge_time_precision": pl.String,
    "title": pl.String,
    "text": pl.String,
    "word_count": pl.Int32,
    "html_url": pl.String,
    "pdf_url": pl.String,
    "language": pl.String,
    "content_hash": pl.String,
    **PROVENANCE,
}

SEP_SCHEMA: dict[str, pl.DataType] = {
    "sep_id": pl.String,
    "meeting_id": pl.String,
    "release_at": UTC_DATETIME,
    "knowledge_at": UTC_DATETIME,
    "knowledge_date": UTC_DATETIME,
    "variable": pl.String,
    "projection_year": pl.String,
    "central_tendency_low": pl.Float64,
    "central_tendency_high": pl.Float64,
    "range_low": pl.Float64,
    "range_high": pl.Float64,
    "median": pl.Float64,
    "unit": pl.String,
    "source_url": pl.String,
    # Which of the accessible version's two layouts the numbers came from.
    "extraction_method": pl.String,
    "parser_version": pl.String,
    "ingested_at": UTC_DATETIME,
}

DOT_PLOT_SCHEMA: dict[str, pl.DataType] = {
    "dot_id": pl.String,
    "meeting_id": pl.String,
    "release_at": UTC_DATETIME,
    "knowledge_at": UTC_DATETIME,
    "knowledge_date": UTC_DATETIME,
    "projection_year": pl.String,
    "rate_level": pl.Float64,
    "dot_count": pl.Int32,
    "source_url": pl.String,
    # Where the dots came from. The Federal Reserve publishes the chart's
    # underlying counts as a table in the accessible version of the
    # projections, so nothing here is read off an image.
    "extraction_method": pl.String,
    "parser_version": pl.String,
    "ingested_at": UTC_DATETIME,
}

VOTE_SCHEMA: dict[str, pl.DataType] = {
    "vote_id": pl.String,
    "meeting_id": pl.String,
    "event_id": pl.String,
    "member_name": pl.String,
    "vote": pl.String,
    "dissent": pl.Boolean,
    "preferred_action_text": pl.String,
    "source_url": pl.String,
    "knowledge_at": UTC_DATETIME,
    "knowledge_date": UTC_DATETIME,
    "parser_version": pl.String,
    "ingested_at": UTC_DATETIME,
}

CALENDAR_SCHEMA: dict[str, pl.DataType] = {
    "calendar_id": pl.String,
    "meeting_id": pl.String,
    "meeting_start_date": pl.Date,
    "meeting_end_date": pl.Date,
    "scheduled_at": UTC_DATETIME,
    "status": pl.String,
    "has_sep_scheduled": pl.Boolean,
    "timezone": pl.String,
    "knowledge_time_precision": pl.String,
    # When the schedule entry itself became knowable: a calendar is a forecast
    # of a publication, never the date it points at.
    "knowledge_date": UTC_DATETIME,
    "source_url": pl.String,
    "ingested_at": UTC_DATETIME,
}

# One row per public number the Committee produced, in the shape ziplime mounts
# a point-in-time bundle from: an entity, the period it is about, and the
# instant it became knowable. Everything else in this dataset is context for
# these rows.
PIT_SCHEMA: dict[str, pl.DataType] = {
    "entity_id": pl.String,
    "event_date": pl.Date,
    "knowledge_date": UTC_DATETIME,
    "meeting_id": pl.String,
    "event_id": pl.String,
    "event_type": pl.String,
    "actual": pl.Float64,
    "previous": pl.Float64,
    "change": pl.Float64,
    # Always null. Consensus forecasts are licensed data and this dataset
    # carries none; the column exists so a bundle can be mounted without one.
    "forecast": pl.Float64,
    "unit": pl.String,
    "decision": pl.String,
    "vote_for": pl.Int32,
    "vote_against": pl.Int32,
    # `reported` where the Committee printed the number, `derived` where this
    # recipe computed it from numbers the Committee printed.
    "value_method": pl.String,
    "knowledge_time_precision": pl.String,
    "source_url": pl.String,
    "parser_version": pl.String,
    "ingested_at": UTC_DATETIME,
}

CONFIG_SCHEMAS: dict[str, dict[str, pl.DataType]] = {
    "events": EVENT_SCHEMA,
    "meetings": MEETING_SCHEMA,
    "documents": DOCUMENT_SCHEMA,
    "sep": SEP_SCHEMA,
    "dot_plot": DOT_PLOT_SCHEMA,
    "votes": VOTE_SCHEMA,
    "calendar": CALENDAR_SCHEMA,
    "pit": PIT_SCHEMA,
}


def empty_frame(schema: Mapping[str, pl.DataType]) -> pl.DataFrame:
    return pl.DataFrame(schema=dict(schema))


def align(frame: pl.DataFrame, schema: Mapping[str, pl.DataType]) -> pl.DataFrame:
    missing = [
        pl.lit(None, dtype=dtype).alias(name)
        for name, dtype in schema.items()
        if name not in frame.columns
    ]
    if missing:
        frame = frame.with_columns(missing)
    return frame.select([pl.col(n).cast(d, strict=False) for n, d in schema.items()])


__all__ = [
    "BUILD_VERSION",
    "CALENDAR_SCHEMA",
    "CONFIG_SCHEMAS",
    "DECISIONS",
    "DOCUMENT_SCHEMA",
    "DOCUMENT_TYPES",
    "DOT_PLOT_SCHEMA",
    "EVENT_SCHEMA",
    "EVENT_TYPES",
    "KNOWLEDGE_PRECISIONS",
    "MEETING_SCHEMA",
    "MEETING_TYPES",
    "PARSER_VERSION",
    "PIT_SCHEMA",
    "PROVENANCE",
    "SEP_SCHEMA",
    "STATUSES",
    "UTC_DATETIME",
    "VOTES",
    "VOTE_SCHEMA",
    "align",
    "empty_frame",
]