Merge branch 'develop' into codex/point

This commit is contained in:
yuxuanhui
2026-08-31 16:14:35 +08:00
86 changed files with 13384 additions and 217 deletions
@@ -0,0 +1,234 @@
"""Create replayable independent sector radar tables."""
from collections.abc import Sequence
import sqlalchemy as sa
from alembic import op
from sqlalchemy.dialects.postgresql import JSONB
revision: str = "0004_sector_radar"
down_revision: str | None = "0003_market_integrity_checks"
branch_labels: str | Sequence[str] | None = None
depends_on: str | Sequence[str] | None = None
def upgrade() -> None:
"""Create source, point-in-time fact, publication, and ranking tables."""
op.create_table(
"sector_radar_source_snapshot",
sa.Column("id", sa.String(64), primary_key=True),
sa.Column("api_name", sa.String(32), nullable=False),
sa.Column("normalized_params", JSONB, nullable=False),
sa.Column("target_trade_date", sa.Date()),
sa.Column("partition_key", sa.String(64)),
sa.Column("observed_at", sa.DateTime(timezone=True), nullable=False),
sa.Column("payload", JSONB, nullable=False),
sa.Column("row_count", sa.Integer(), nullable=False),
sa.Column("returned_fields", JSONB, nullable=False),
sa.Column("content_sha256", sa.String(64), nullable=False),
sa.Column("row_limit", sa.Integer()),
sa.Column("limit_reached", sa.Boolean(), nullable=False),
sa.Column(
"created_at",
sa.DateTime(timezone=True),
nullable=False,
server_default=sa.text("now()"),
),
sa.CheckConstraint("row_count >= 0", name="ck_sector_radar_source_row_count"),
sa.CheckConstraint(
"row_limit IS NULL OR row_limit > 0", name="ck_sector_radar_source_limit"
),
)
op.create_index(
"ix_sector_radar_source_api_date",
"sector_radar_source_snapshot",
["api_name", "target_trade_date", "observed_at"],
)
op.create_table(
"sector_radar_membership",
sa.Column(
"source_snapshot_id",
sa.String(64),
sa.ForeignKey("sector_radar_source_snapshot.id", ondelete="CASCADE"),
nullable=False,
),
sa.Column("trade_date", sa.Date(), nullable=False),
sa.Column("sector_type", sa.String(16), nullable=False),
sa.Column("sector_code", sa.String(16), nullable=False),
sa.Column("sector_name", sa.String(128), nullable=False),
sa.Column("stock_code", sa.String(12), nullable=False),
sa.Column("stock_name", sa.String(128), nullable=False),
sa.Column("membership_status", sa.String(32), nullable=False),
sa.Column(
"created_at",
sa.DateTime(timezone=True),
nullable=False,
server_default=sa.text("now()"),
),
sa.PrimaryKeyConstraint("source_snapshot_id", "sector_code", "stock_code"),
sa.CheckConstraint(
"sector_type IN ('concept', 'industry')",
name="ck_sector_radar_membership_type",
),
sa.CheckConstraint(
"membership_status = 'available'",
name="ck_sector_radar_membership_status",
),
)
op.create_index(
"ix_sector_radar_membership_date_sector",
"sector_radar_membership",
["trade_date", "sector_type", "sector_code"],
)
op.create_table(
"sector_radar_stock_fact",
sa.Column("fact_revision", sa.String(64), nullable=False),
sa.Column("trade_date", sa.Date(), nullable=False),
sa.Column("ts_code", sa.String(12), nullable=False),
sa.Column("source_snapshot_ids", JSONB, nullable=False),
sa.Column("status", sa.String(32), nullable=False),
sa.Column("turnover_yuan", sa.Numeric(28, 6)),
sa.Column("net_amount_yuan", sa.Numeric(28, 6)),
sa.Column(
"created_at",
sa.DateTime(timezone=True),
nullable=False,
server_default=sa.text("now()"),
),
sa.PrimaryKeyConstraint("fact_revision", "ts_code"),
sa.CheckConstraint(
"turnover_yuan IS NULL OR turnover_yuan >= 0",
name="ck_sector_radar_stock_turnover",
),
)
op.create_index(
"ix_sector_radar_stock_fact_date",
"sector_radar_stock_fact",
["trade_date", "ts_code"],
)
op.create_table(
"sector_radar_publication",
sa.Column("id", sa.String(64), primary_key=True),
sa.Column("target_trade_date", sa.Date(), nullable=False),
sa.Column("status", sa.String(16), nullable=False),
sa.Column("source_version", sa.String(128), nullable=False),
sa.Column("universe_version", sa.String(128), nullable=False),
sa.Column("metric_versions", JSONB, nullable=False),
sa.Column("input_hash", sa.String(64)),
sa.Column("coverage", sa.Numeric(8, 6), nullable=False),
sa.Column("started_at", sa.DateTime(timezone=True), nullable=False),
sa.Column("finished_at", sa.DateTime(timezone=True)),
sa.Column("error_summary", sa.String(500)),
sa.Column(
"created_at",
sa.DateTime(timezone=True),
nullable=False,
server_default=sa.text("now()"),
),
sa.CheckConstraint(
"status IN ('running', 'success', 'partial', 'failed')",
name="ck_sector_radar_publication_status",
),
sa.CheckConstraint(
"coverage >= 0 AND coverage <= 1",
name="ck_sector_radar_publication_coverage",
),
sa.CheckConstraint(
"(status = 'running' AND finished_at IS NULL) OR "
"(status <> 'running' AND finished_at IS NOT NULL)",
name="ck_sector_radar_publication_finished",
),
)
op.create_index(
"ix_sector_radar_publication_status_date",
"sector_radar_publication",
["status", "target_trade_date", "finished_at"],
)
op.create_index(
"uq_sector_radar_publication_running_date",
"sector_radar_publication",
["target_trade_date"],
unique=True,
postgresql_where=sa.text("status = 'running'"),
)
op.create_table(
"sector_radar_ranking",
sa.Column(
"publication_id",
sa.String(64),
sa.ForeignKey("sector_radar_publication.id", ondelete="CASCADE"),
nullable=False,
),
sa.Column("trade_date", sa.Date(), nullable=False),
sa.Column("sector_type", sa.String(16), nullable=False),
sa.Column("sector_code", sa.String(16), nullable=False),
sa.Column("sector_name", sa.String(128), nullable=False),
sa.Column("metric_kind", sa.String(16), nullable=False),
sa.Column("metric_version", sa.String(128), nullable=False),
sa.Column("implementation_kind", sa.String(16), nullable=False),
sa.Column("unit", sa.String(16), nullable=False),
sa.Column("metric_value", sa.Numeric(28, 12)),
sa.Column("quality", sa.String(32), nullable=False),
sa.Column("member_count", sa.Integer(), nullable=False),
sa.Column("valid_sample_count", sa.Integer(), nullable=False),
sa.Column("membership_coverage", sa.Numeric(8, 6), nullable=False),
sa.Column("moneyflow_coverage", sa.Numeric(8, 6), nullable=False),
sa.Column("rank_position", sa.Integer()),
sa.Column("rank_percentile", sa.Numeric(18, 12)),
sa.Column("rank_changes", JSONB, nullable=False),
sa.Column(
"created_at",
sa.DateTime(timezone=True),
nullable=False,
server_default=sa.text("now()"),
),
sa.PrimaryKeyConstraint(
"publication_id",
"sector_type",
"sector_code",
"metric_version",
),
sa.CheckConstraint(
"sector_type IN ('concept', 'industry')",
name="ck_sector_radar_ranking_type",
),
sa.CheckConstraint(
"implementation_kind = 'independent'",
name="ck_sector_radar_ranking_implementation",
),
)
op.create_index(
"ix_sector_radar_ranking_query",
"sector_radar_ranking",
["publication_id", "sector_type", "metric_version", "rank_position"],
)
def downgrade() -> None:
"""Drop only sector radar tables in dependency-safe order."""
op.drop_index("ix_sector_radar_ranking_query", table_name="sector_radar_ranking")
op.drop_table("sector_radar_ranking")
op.drop_index(
"uq_sector_radar_publication_running_date",
table_name="sector_radar_publication",
)
op.drop_index(
"ix_sector_radar_publication_status_date",
table_name="sector_radar_publication",
)
op.drop_table("sector_radar_publication")
op.drop_index("ix_sector_radar_stock_fact_date", table_name="sector_radar_stock_fact")
op.drop_table("sector_radar_stock_fact")
op.drop_index(
"ix_sector_radar_membership_date_sector",
table_name="sector_radar_membership",
)
op.drop_table("sector_radar_membership")
op.drop_index("ix_sector_radar_source_api_date", table_name="sector_radar_source_snapshot")
op.drop_table("sector_radar_source_snapshot")
@@ -0,0 +1,155 @@
"""Persist publication-owned sector daily aggregates for exact metric replay."""
from collections.abc import Sequence
import sqlalchemy as sa
from alembic import op
revision: str = "0005_radar_daily_aggregate"
down_revision: str | None = "0004_sector_radar"
branch_labels: str | Sequence[str] | None = None
depends_on: str | Sequence[str] | None = None
def upgrade() -> None:
"""Create source recovery links and exact multi-day metric inputs."""
op.create_table(
"sector_radar_publication_source",
sa.Column(
"publication_id",
sa.String(64),
sa.ForeignKey("sector_radar_publication.id", ondelete="CASCADE"),
nullable=False,
),
sa.Column("source_group", sa.String(32), nullable=False),
sa.Column("source_order", sa.Integer(), nullable=False),
sa.Column(
"refresh_on_retry",
sa.Boolean(),
nullable=False,
server_default=sa.false(),
),
sa.Column(
"source_snapshot_id",
sa.String(64),
sa.ForeignKey("sector_radar_source_snapshot.id", ondelete="RESTRICT"),
nullable=False,
),
sa.PrimaryKeyConstraint("publication_id", "source_group", "source_order"),
sa.UniqueConstraint(
"publication_id",
"source_group",
"source_snapshot_id",
name="uq_sector_radar_publication_source_snapshot",
),
sa.CheckConstraint(
"source_group IN ('calendar', 'concept_indices', 'industry_indices', "
"'members', 'stock_basics', 'suspensions', 'daily', 'moneyflow_dc')",
name="ck_sector_radar_publication_source_group",
),
sa.CheckConstraint(
"source_order >= 0",
name="ck_sector_radar_publication_source_order",
),
)
op.create_index(
"ix_sector_radar_publication_source_snapshot",
"sector_radar_publication_source",
["source_snapshot_id"],
)
op.create_unique_constraint(
"uq_sector_radar_publication_id_date",
"sector_radar_publication",
["id", "target_trade_date"],
)
op.create_foreign_key(
"fk_sector_radar_ranking_publication_date",
"sector_radar_ranking",
"sector_radar_publication",
["publication_id", "trade_date"],
["id", "target_trade_date"],
ondelete="CASCADE",
)
op.create_table(
"sector_radar_daily_aggregate",
sa.Column("publication_id", sa.String(64), nullable=False),
sa.Column("trade_date", sa.Date(), nullable=False),
sa.Column("sector_type", sa.String(16), nullable=False),
sa.Column("sector_code", sa.String(16), nullable=False),
sa.Column("sector_name", sa.String(128), nullable=False),
sa.Column("member_count", sa.Integer(), nullable=False),
sa.Column("valid_sample_count", sa.Integer(), nullable=False),
sa.Column("net_amount_yuan", sa.Numeric(28, 6)),
sa.Column("turnover_yuan", sa.Numeric(28, 6)),
sa.Column("membership_coverage", sa.Numeric(8, 6), nullable=False),
sa.Column("moneyflow_coverage", sa.Numeric(8, 6), nullable=False),
sa.Column(
"created_at",
sa.DateTime(timezone=True),
nullable=False,
server_default=sa.text("now()"),
),
sa.PrimaryKeyConstraint("publication_id", "sector_type", "sector_code"),
sa.ForeignKeyConstraint(
["publication_id", "trade_date"],
["sector_radar_publication.id", "sector_radar_publication.target_trade_date"],
name="fk_sector_radar_daily_aggregate_publication_date",
ondelete="CASCADE",
),
sa.CheckConstraint(
"sector_type IN ('concept', 'industry')",
name="ck_sector_radar_daily_aggregate_type",
),
sa.CheckConstraint(
"member_count >= 0 AND valid_sample_count >= 0 AND valid_sample_count <= member_count",
name="ck_sector_radar_daily_aggregate_counts",
),
sa.CheckConstraint(
"membership_coverage >= 0 AND membership_coverage <= 1 "
"AND moneyflow_coverage >= 0 AND moneyflow_coverage <= 1",
name="ck_sector_radar_daily_aggregate_coverage",
),
sa.CheckConstraint(
"turnover_yuan IS NULL OR (turnover_yuan >= 0 AND "
"turnover_yuan NOT IN ('NaN'::numeric, 'Infinity'::numeric))",
name="ck_sector_radar_daily_aggregate_turnover",
),
sa.CheckConstraint(
"net_amount_yuan IS NULL OR net_amount_yuan NOT IN "
"('NaN'::numeric, 'Infinity'::numeric, '-Infinity'::numeric)",
name="ck_sector_radar_daily_aggregate_net_amount",
),
)
op.create_index(
"ix_sector_radar_daily_aggregate_history",
"sector_radar_daily_aggregate",
["trade_date", "sector_type", "sector_code"],
)
def downgrade() -> None:
"""Drop only the replay aggregate extension."""
op.drop_index(
"ix_sector_radar_daily_aggregate_history",
table_name="sector_radar_daily_aggregate",
)
op.drop_table("sector_radar_daily_aggregate")
op.drop_constraint(
"fk_sector_radar_ranking_publication_date",
"sector_radar_ranking",
type_="foreignkey",
)
op.drop_constraint(
"uq_sector_radar_publication_id_date",
"sector_radar_publication",
type_="unique",
)
op.drop_index(
"ix_sector_radar_publication_source_snapshot",
table_name="sector_radar_publication_source",
)
op.drop_table("sector_radar_publication_source")
@@ -0,0 +1,103 @@
"""Persist explicit unknown point-in-time sector membership snapshots."""
from collections.abc import Sequence
import sqlalchemy as sa
from alembic import op
revision: str = "0006_membership_unknown"
down_revision: str | None = "0005_radar_daily_aggregate"
branch_labels: str | Sequence[str] | None = None
depends_on: str | Sequence[str] | None = None
def upgrade() -> None:
"""Allow one null-stock marker for an explicitly empty sector partition."""
op.add_column(
"sector_radar_membership",
sa.Column("membership_key", sa.String(32), nullable=True),
)
op.execute("UPDATE sector_radar_membership SET membership_key = stock_code")
op.drop_constraint(
"sector_radar_membership_pkey",
"sector_radar_membership",
type_="primary",
)
op.drop_constraint(
"ck_sector_radar_membership_status",
"sector_radar_membership",
type_="check",
)
op.alter_column(
"sector_radar_membership",
"membership_key",
existing_type=sa.String(32),
nullable=False,
)
op.alter_column(
"sector_radar_membership",
"stock_code",
existing_type=sa.String(12),
nullable=True,
)
op.alter_column(
"sector_radar_membership",
"stock_name",
existing_type=sa.String(128),
nullable=True,
)
op.create_primary_key(
"sector_radar_membership_pkey",
"sector_radar_membership",
["source_snapshot_id", "sector_code", "membership_key"],
)
op.create_check_constraint(
"ck_sector_radar_membership_status",
"sector_radar_membership",
"(membership_status = 'available' "
"AND stock_code IS NOT NULL AND stock_name IS NOT NULL "
"AND membership_key = stock_code) OR "
"(membership_status = 'membership_unknown' "
"AND stock_code IS NULL AND stock_name IS NULL "
"AND membership_key = '__membership_unknown__')",
)
def downgrade() -> None:
"""Discard unknown markers and restore the available-member-only schema."""
op.execute("DELETE FROM sector_radar_membership WHERE membership_status = 'membership_unknown'")
op.drop_constraint(
"ck_sector_radar_membership_status",
"sector_radar_membership",
type_="check",
)
op.drop_constraint(
"sector_radar_membership_pkey",
"sector_radar_membership",
type_="primary",
)
op.alter_column(
"sector_radar_membership",
"stock_code",
existing_type=sa.String(12),
nullable=False,
)
op.alter_column(
"sector_radar_membership",
"stock_name",
existing_type=sa.String(128),
nullable=False,
)
op.drop_column("sector_radar_membership", "membership_key")
op.create_primary_key(
"sector_radar_membership_pkey",
"sector_radar_membership",
["source_snapshot_id", "sector_code", "stock_code"],
)
op.create_check_constraint(
"ck_sector_radar_membership_status",
"sector_radar_membership",
"membership_status = 'available'",
)