sediment-core 0.1.0__py3-none-any.whl

sediment_core/__init__.py

@@ -0,0 +1,144 @@
+# SPDX-License-Identifier: AGPL-3.0-or-later
+"""Sediment core: fact models and the PostgreSQL fact store."""
+
+from .models import (
+    BranchName,
+    CIOutcome,
+    CIReason,
+    CI_REASON_MAX_LENGTH,
+    CIProvider,
+    CIResult,
+    CommitSha,
+    AgentHarness,
+    DeveloperDecision,
+    EditObservation,
+    RejectedEdit,
+    RetryLinkage,
+    InteractionMode,
+    FactTable,
+    ForgeProvider,
+    ForgeHost,
+    ProviderRepositoryId,
+    RepositoryRename,
+    GatewayProvider,
+    InferenceCall,
+    InferenceMessage,
+    InferenceMessagePart,
+    ModelName,
+    NonEmptyId,
+    NonEmptyContent,
+    ScalarIdentity,
+    OrgId,
+    Push,
+    SessionCommitObservation,
+    PullRequestMerge,
+    PullRequestRevision,
+    QuarantineAction,
+    QuarantineRecord,
+    ReasoningPart,
+    RepoSlug,
+    RequiredRepoSlug,
+    ToolCallPart,
+    ToolCallResponsePart,
+    TextPart,
+    UNKNOWN_USER_ID,
+    WorkflowName,
+    WorkflowPath,
+    normalize_commit_sha,
+    normalize_repo_slug,
+)
+from .org import normalize_org_id
+from .redaction import REDACTION_MARKER, RedactionReason, redact_fact
+from .store import (
+    COMPOSITE_FILTER_KEY_LIMIT,
+    FactStore,
+    InferenceCallIdentityConflict,
+    InferenceCallReceipt,
+    RepositoryFactReceipt,
+    RepositoryIdentityConflict,
+    RepositoryIdentityEvidence,
+    RepositoryReadAmbiguous,
+    RepositoryReadKey,
+    RepositoryCommitReadKey,
+    RepositoryPRReadKey,
+    REPOSITORY_FILTER_KEY_LIMIT,
+    REPOSITORY_IDENTITY_LIMIT,
+    OperationalReportLimitExceeded,
+    ReportInferenceCall,
+    INFERENCE_CALL_ROW_BYTES_LIMIT,
+    INFERENCE_REPORT_ROW_LIMIT,
+    INFERENCE_SESSION_BYTES_LIMIT,
+    INFERENCE_PROJECTION_BYTES_LIMIT,
+    validate_composite_filter_keys,
+)
+
+__all__ = [
+    "BranchName",
+    "CIOutcome",
+    "CIReason",
+    "CI_REASON_MAX_LENGTH",
+    "CIProvider",
+    "CIResult",
+    "COMPOSITE_FILTER_KEY_LIMIT",
+    "CommitSha",
+    "AgentHarness",
+    "DeveloperDecision",
+    "EditObservation",
+    "RejectedEdit",
+    "RetryLinkage",
+    "InteractionMode",
+    "FactStore",
+    "FactTable",
+    "ForgeProvider",
+    "ForgeHost",
+    "ProviderRepositoryId",
+    "RepositoryRename",
+    "GatewayProvider",
+    "InferenceCall",
+    "InferenceCallIdentityConflict",
+    "InferenceCallReceipt",
+    "RepositoryFactReceipt",
+    "RepositoryIdentityConflict",
+    "RepositoryIdentityEvidence",
+    "RepositoryReadAmbiguous",
+    "RepositoryReadKey",
+    "RepositoryCommitReadKey",
+    "RepositoryPRReadKey",
+    "REPOSITORY_FILTER_KEY_LIMIT",
+    "REPOSITORY_IDENTITY_LIMIT",
+    "InferenceMessage",
+    "InferenceMessagePart",
+    "ModelName",
+    "NonEmptyId",
+    "NonEmptyContent",
+    "ScalarIdentity",
+    "OperationalReportLimitExceeded",
+    "ReportInferenceCall",
+    "INFERENCE_CALL_ROW_BYTES_LIMIT",
+    "INFERENCE_REPORT_ROW_LIMIT",
+    "INFERENCE_SESSION_BYTES_LIMIT",
+    "INFERENCE_PROJECTION_BYTES_LIMIT",
+    "OrgId",
+    "Push",
+    "SessionCommitObservation",
+    "PullRequestMerge",
+    "PullRequestRevision",
+    "QuarantineAction",
+    "QuarantineRecord",
+    "REDACTION_MARKER",
+    "RedactionReason",
+    "ReasoningPart",
+    "RepoSlug",
+    "RequiredRepoSlug",
+    "ToolCallPart",
+    "ToolCallResponsePart",
+    "TextPart",
+    "UNKNOWN_USER_ID",
+    "WorkflowName",
+    "WorkflowPath",
+    "normalize_commit_sha",
+    "normalize_org_id",
+    "normalize_repo_slug",
+    "redact_fact",
+    "validate_composite_filter_keys",
+]

sediment_core/alembic/__init__.py

@@ -0,0 +1,2 @@
+# SPDX-License-Identifier: AGPL-3.0-or-later
+"""Alembic migration environment for the PostgreSQL fact store."""

sediment_core/alembic/env.py

@@ -0,0 +1,25 @@
+# SPDX-License-Identifier: AGPL-3.0-or-later
+"""Run Alembic migrations on the connection supplied by Sediment."""
+
+from __future__ import annotations
+
+from alembic import context
+
+from sediment_core.postgres_schema import metadata
+
+
+def run_migrations() -> None:
+    connection = context.config.attributes.get("connection")
+    if connection is None:
+        raise RuntimeError("Sediment must supply an Alembic connection")
+    context.configure(
+        connection=connection,
+        target_metadata=metadata,
+        compare_type=True,
+        transaction_per_migration=True,
+    )
+    with context.begin_transaction():
+        context.run_migrations()
+
+
+run_migrations()

sediment_core/alembic/versions/0001_postgresql_baseline.py

@@ -0,0 +1,500 @@
+# SPDX-License-Identifier: AGPL-3.0-or-later
+"""Create the frozen PostgreSQL fact-store baseline.
+
+Revision ID: 0001_postgresql_baseline
+Revises:
+"""
+
+from __future__ import annotations
+
+from alembic import op
+import sqlalchemy as sa
+
+revision = "0001_postgresql_baseline"
+down_revision = None
+branch_labels = None
+depends_on = None
+
+
+def upgrade() -> None:
+    # ### commands auto generated by Alembic - please adjust! ###
+    op.create_table(
+        "ci_outcomes",
+        sa.Column("schema_version", sa.BigInteger(), nullable=False),
+        sa.Column("outcome_id", sa.Text(), nullable=False),
+        sa.Column("org_id", sa.Text(), nullable=False),
+        sa.Column("provider", sa.Text(), nullable=False),
+        sa.Column("run_id", sa.Text(), nullable=False),
+        sa.Column("run_attempt", sa.BigInteger(), nullable=True),
+        sa.Column("repo", sa.Text(), nullable=False),
+        sa.Column("commit_sha", sa.Text(), nullable=False),
+        sa.Column("branch", sa.Text(), nullable=False),
+        sa.Column("result", sa.Text(), nullable=False),
+        sa.Column("workflow_name", sa.Text(), nullable=False),
+        sa.Column("workflow_id", sa.Text(), nullable=True),
+        sa.Column("workflow_path", sa.Text(), nullable=True),
+        sa.Column("run_url", sa.Text(), nullable=True),
+        sa.Column("provider_result", sa.Text(), nullable=True),
+        sa.Column("error_type", sa.Text(), nullable=True),
+        sa.Column("reason", sa.Text(), nullable=True),
+        sa.Column("source_event_type", sa.Text(), nullable=True),
+        sa.Column("source_spec_version", sa.Text(), nullable=True),
+        sa.Column("source_event_id", sa.Text(), nullable=True),
+        sa.Column("pr_number", sa.BigInteger(), nullable=True),
+        sa.Column("captured_at", sa.DateTime(timezone=True), nullable=False),
+        sa.Column("raw", sa.Text(), nullable=False),
+        sa.CheckConstraint(
+            "branch NOT LIKE 'refs/heads/%' AND branch = btrim(branch, E' \\t\\n\\r')",
+            name="ck_ci_outcomes_branch",
+        ),
+        sa.CheckConstraint(
+            "commit_sha ~ '^(?:[0-9a-f]{40}|[0-9a-f]{64})$'",
+            name="ck_ci_outcomes_commit_sha",
+        ),
+        sa.CheckConstraint(
+            "error_type IS NULL OR (error_type <> '' AND error_type = btrim(error_type, E' \\t\\n\\r'))",
+            name="ck_ci_outcomes_error_type",
+        ),
+        sa.CheckConstraint(
+            "org_id ~ '^[a-z0-9][a-z0-9._-]{0,63}$'", name="ck_ci_outcomes_org_id"
+        ),
+        sa.CheckConstraint(
+            "outcome_id <> '' AND outcome_id = btrim(outcome_id, E' \\t\\n\\r')",
+            name="ck_ci_outcomes_outcome_id",
+        ),
+        sa.CheckConstraint(
+            "provider IN ('github_actions', 'jenkins', 'gitlab_ci', 'circleci', 'buildkite', 'other')",
+            name="ck_ci_outcomes_provider",
+        ),
+        sa.CheckConstraint(
+            "provider_result IS NULL OR (provider_result <> '' AND provider_result = btrim(provider_result, E' \\t\\n\\r'))",
+            name="ck_ci_outcomes_provider_result",
+        ),
+        sa.CheckConstraint(
+            "repo = '' OR (repo = lower(repo) AND repo ~ '^[^/]+/[^/]+$')",
+            name="ck_ci_outcomes_repo",
+        ),
+        sa.CheckConstraint(
+            "result IN ('passed', 'failed', 'error', 'timed_out', 'cancelled', 'skipped', 'neutral', 'unknown')",
+            name="ck_ci_outcomes_result",
+        ),
+        sa.CheckConstraint(
+            "run_id <> '' AND run_id = btrim(run_id, E' \\t\\n\\r')",
+            name="ck_ci_outcomes_run_id",
+        ),
+        sa.CheckConstraint(
+            "run_url IS NULL OR (run_url <> '' AND run_url = btrim(run_url, E' \\t\\n\\r'))",
+            name="ck_ci_outcomes_run_url",
+        ),
+        sa.CheckConstraint(
+            "source_event_id IS NULL OR (source_event_id <> '' AND source_event_id = btrim(source_event_id, E' \\t\\n\\r'))",
+            name="ck_ci_outcomes_source_event_id",
+        ),
+        sa.CheckConstraint(
+            "source_event_type IS NULL OR (source_event_type <> '' AND source_event_type = btrim(source_event_type, E' \\t\\n\\r'))",
+            name="ck_ci_outcomes_source_event_type",
+        ),
+        sa.CheckConstraint(
+            "source_spec_version IS NULL OR (source_spec_version <> '' AND source_spec_version = btrim(source_spec_version, E' \\t\\n\\r'))",
+            name="ck_ci_outcomes_source_spec_version",
+        ),
+        sa.CheckConstraint(
+            "workflow_id IS NULL OR (workflow_id <> '' AND workflow_id = btrim(workflow_id, E' \\t\\n\\r'))",
+            name="ck_ci_outcomes_workflow_id",
+        ),
+        sa.CheckConstraint(
+            "pr_number IS NULL OR pr_number BETWEEN 1 AND 9223372036854775807",
+            name="ck_ci_outcomes_pr_number",
+        ),
+        sa.CheckConstraint(
+            "reason IS NULL OR char_length(reason) <= 4096",
+            name="ck_ci_outcomes_reason",
+        ),
+        sa.CheckConstraint(
+            "run_attempt IS NULL OR run_attempt BETWEEN 1 AND 9223372036854775807",
+            name="ck_ci_outcomes_run_attempt",
+        ),
+        sa.CheckConstraint("schema_version = 1", name="ck_ci_outcomes_schema_version"),
+        sa.PrimaryKeyConstraint("outcome_id"),
+    )
+    op.create_index(
+        "ix_ci_org_commit", "ci_outcomes", ["org_id", "commit_sha"], unique=False
+    )
+    op.create_index(
+        "uq_ci_run",
+        "ci_outcomes",
+        ["org_id", "provider", "run_id", sa.literal_column("coalesce(run_attempt, 0)")],
+        unique=True,
+    )
+    op.create_table(
+        "developer_decisions",
+        sa.Column("decision_id", sa.Text(), nullable=False),
+        sa.Column("org_id", sa.Text(), nullable=False),
+        sa.Column("session_id", sa.Text(), nullable=False),
+        sa.Column("user_id", sa.Text(), nullable=True),
+        sa.Column("agent_harness", sa.Text(), nullable=False),
+        sa.Column("file_path", sa.Text(), nullable=False),
+        sa.Column("accepted", sa.Boolean(), nullable=False),
+        sa.Column("explicit", sa.Boolean(), nullable=False),
+        sa.Column("interaction_mode", sa.Text(), nullable=False),
+        sa.Column("commit_sha", sa.Text(), nullable=True),
+        sa.Column("call_id", sa.Text(), nullable=True),
+        sa.Column("edit_retention_score", sa.Float(), nullable=True),
+        sa.Column("observation_delay_ms", sa.BigInteger(), nullable=True),
+        sa.Column("occurred_at", sa.DateTime(timezone=True), nullable=False),
+        sa.Column("captured_at", sa.DateTime(timezone=True), nullable=False),
+        sa.Column("raw", sa.Text(), nullable=False),
+        sa.CheckConstraint(
+            "agent_harness IN ('claude-code', 'copilot', 'codex', 'pi')",
+            name="ck_decisions_agent_harness",
+        ),
+        sa.CheckConstraint(
+            "call_id IS NULL OR (call_id <> '' AND call_id = btrim(call_id, E' \\t\\n\\r'))",
+            name="ck_decisions_call_id",
+        ),
+        sa.CheckConstraint(
+            "commit_sha IS NULL OR (commit_sha ~ '^(?:[0-9a-f]{40}|[0-9a-f]{64})$')",
+            name="ck_decisions_commit_sha",
+        ),
+        sa.CheckConstraint(
+            "decision_id <> '' AND decision_id = btrim(decision_id, E' \\t\\n\\r')",
+            name="ck_decisions_decision_id",
+        ),
+        sa.CheckConstraint(
+            "interaction_mode IN ('agent', 'inline')",
+            name="ck_decisions_interaction_mode",
+        ),
+        sa.CheckConstraint(
+            "org_id ~ '^[a-z0-9][a-z0-9._-]{0,63}$'", name="ck_decisions_org_id"
+        ),
+        sa.CheckConstraint(
+            "session_id <> '' AND session_id = btrim(session_id, E' \\t\\n\\r')",
+            name="ck_decisions_session_id",
+        ),
+        sa.CheckConstraint(
+            "user_id IS NULL OR (user_id <> '' AND user_id = btrim(user_id, E' \\t\\n\\r'))",
+            name="ck_decisions_user_id",
+        ),
+        sa.CheckConstraint(
+            "edit_retention_score IS NULL OR edit_retention_score BETWEEN 0.0 AND 1.0",
+            name="ck_decisions_edit_retention_score",
+        ),
+        sa.CheckConstraint(
+            "observation_delay_ms IS NULL OR observation_delay_ms >= 0",
+            name="ck_decisions_observation_delay_ms",
+        ),
+        sa.PrimaryKeyConstraint("decision_id"),
+    )
+    op.create_index(
+        "uq_decisions_keyed",
+        "developer_decisions",
+        [
+            "org_id",
+            "agent_harness",
+            "session_id",
+            "accepted",
+            "explicit",
+            "interaction_mode",
+            "occurred_at",
+            "call_id",
+        ],
+        unique=True,
+        postgresql_where=sa.text(
+            "call_id IS NOT NULL AND agent_harness = 'claude-code'"
+        ),
+    )
+    op.create_index(
+        "uq_decisions_natural",
+        "developer_decisions",
+        [
+            "org_id",
+            "agent_harness",
+            "session_id",
+            "file_path",
+            "accepted",
+            "explicit",
+            "interaction_mode",
+            "occurred_at",
+            sa.literal_column("coalesce(call_id, '')"),
+            sa.literal_column("coalesce(observation_delay_ms, -1)"),
+        ],
+        unique=True,
+        postgresql_where=sa.text("call_id IS NULL OR agent_harness != 'claude-code'"),
+    )
+    op.create_table(
+        "edit_observations",
+        sa.Column("observation_id", sa.Text(), nullable=False),
+        sa.Column("org_id", sa.Text(), nullable=False),
+        sa.Column("session_id", sa.Text(), nullable=False),
+        sa.Column("user_id", sa.Text(), nullable=True),
+        sa.Column("agent_harness", sa.Text(), nullable=False),
+        sa.Column("file_path", sa.Text(), nullable=False),
+        sa.Column("call_id", sa.Text(), nullable=False),
+        sa.Column("applied_text", sa.Text(), nullable=False),
+        sa.Column("observed_file_text", sa.Text(), nullable=False),
+        sa.Column("external_lines_added", sa.BigInteger(), nullable=True),
+        sa.Column("external_lines_removed", sa.BigInteger(), nullable=True),
+        sa.Column("occurred_at", sa.DateTime(timezone=True), nullable=False),
+        sa.Column("captured_at", sa.DateTime(timezone=True), nullable=False),
+        sa.Column("raw", sa.Text(), nullable=False),
+        sa.CheckConstraint(
+            "agent_harness IN ('claude-code', 'copilot', 'codex', 'pi')",
+            name="ck_edit_observations_agent_harness",
+        ),
+        sa.CheckConstraint(
+            "call_id <> '' AND call_id = btrim(call_id, E' \\t\\n\\r')",
+            name="ck_edit_observations_call_id",
+        ),
+        sa.CheckConstraint(
+            "observation_id <> '' AND observation_id = btrim(observation_id, E' \\t\\n\\r')",
+            name="ck_edit_observations_observation_id",
+        ),
+        sa.CheckConstraint(
+            "org_id ~ '^[a-z0-9][a-z0-9._-]{0,63}$'", name="ck_edit_observations_org_id"
+        ),
+        sa.CheckConstraint(
+            "session_id <> '' AND session_id = btrim(session_id, E' \\t\\n\\r')",
+            name="ck_edit_observations_session_id",
+        ),
+        sa.CheckConstraint(
+            "user_id IS NULL OR (user_id <> '' AND user_id = btrim(user_id, E' \\t\\n\\r'))",
+            name="ck_edit_observations_user_id",
+        ),
+        sa.CheckConstraint(
+            "(external_lines_added IS NULL OR external_lines_added >= 0) AND (external_lines_removed IS NULL OR external_lines_removed >= 0)",
+            name="ck_edit_observations_external_counts",
+        ),
+        sa.PrimaryKeyConstraint("observation_id"),
+    )
+    op.create_index(
+        "uq_edit_observations_call",
+        "edit_observations",
+        ["org_id", "agent_harness", "session_id", "call_id"],
+        unique=True,
+    )
+    op.create_table(
+        "fact_quarantine",
+        sa.Column(
+            "quarantine_revision",
+            sa.BigInteger(),
+            sa.Identity(always=True),
+            nullable=False,
+        ),
+        sa.Column("quarantine_id", sa.Text(), nullable=False),
+        sa.Column("org_id", sa.Text(), nullable=False),
+        sa.Column("fact_table", sa.Text(), nullable=False),
+        sa.Column("fact_id", sa.Text(), nullable=False),
+        sa.Column("action", sa.Text(), nullable=False),
+        sa.Column("reason", sa.Text(), nullable=False),
+        sa.Column("recorded_at", sa.DateTime(timezone=True), nullable=False),
+        sa.CheckConstraint(
+            "action IN ('quarantine', 'release')", name="ck_quarantine_action"
+        ),
+        sa.CheckConstraint(
+            "btrim(reason, E' \\t\\n\\r') <> ''", name="ck_quarantine_reason"
+        ),
+        sa.CheckConstraint(
+            "fact_id <> '' AND fact_id = btrim(fact_id, E' \\t\\n\\r')",
+            name="ck_quarantine_fact_id",
+        ),
+        sa.CheckConstraint(
+            "fact_table IN ('inference_calls', 'developer_decisions', 'ci_outcomes', 'pushes', 'edit_observations', 'rejected_edits')",
+            name="ck_quarantine_table",
+        ),
+        sa.CheckConstraint(
+            "org_id ~ '^[a-z0-9][a-z0-9._-]{0,63}$'", name="ck_quarantine_org_id"
+        ),
+        sa.CheckConstraint(
+            "quarantine_id <> '' AND quarantine_id = btrim(quarantine_id, E' \\t\\n\\r')",
+            name="ck_quarantine_quarantine_id",
+        ),
+        sa.PrimaryKeyConstraint("quarantine_revision"),
+        sa.UniqueConstraint("quarantine_id", name="uq_quarantine_id"),
+    )
+    op.create_index(
+        "ix_quarantine_fact",
+        "fact_quarantine",
+        ["org_id", "fact_table", "fact_id", "quarantine_revision"],
+        unique=False,
+    )
+    op.create_table(
+        "inference_calls",
+        sa.Column("schema_version", sa.BigInteger(), nullable=False),
+        sa.Column("inference_call_id", sa.Text(), nullable=False),
+        sa.Column("org_id", sa.Text(), nullable=False),
+        sa.Column("session_id", sa.Text(), nullable=False),
+        sa.Column("user_id", sa.Text(), nullable=True),
+        sa.Column("gateway_provider", sa.Text(), nullable=False),
+        sa.Column("model_provider", sa.Text(), nullable=True),
+        sa.Column("model", sa.Text(), nullable=True),
+        sa.Column("input_messages", sa.Text(), nullable=False),
+        sa.Column("output_messages", sa.Text(), nullable=False),
+        sa.Column("input_tokens", sa.BigInteger(), nullable=True),
+        sa.Column("output_tokens", sa.BigInteger(), nullable=True),
+        sa.Column("duration_ms", sa.BigInteger(), nullable=True),
+        sa.Column("model_call_id", sa.Text(), nullable=True),
+        sa.Column("observed_at", sa.DateTime(timezone=True), nullable=False),
+        sa.Column("raw", sa.Text(), nullable=False),
+        sa.CheckConstraint(
+            "gateway_provider IN ('litellm', 'portkey', 'helicone', 'unknown')",
+            name="ck_inference_calls_gateway_provider",
+        ),
+        sa.CheckConstraint(
+            "inference_call_id <> '' AND inference_call_id = btrim(inference_call_id, E' \\t\\n\\r')",
+            name="ck_inference_calls_inference_call_id",
+        ),
+        sa.CheckConstraint(
+            "model_call_id IS NULL OR (model_call_id <> '' AND model_call_id = btrim(model_call_id, E' \\t\\n\\r'))",
+            name="ck_inference_calls_model_call_id",
+        ),
+        sa.CheckConstraint(
+            "org_id ~ '^[a-z0-9][a-z0-9._-]{0,63}$'", name="ck_inference_calls_org_id"
+        ),
+        sa.CheckConstraint(
+            "session_id <> '' AND session_id = btrim(session_id, E' \\t\\n\\r')",
+            name="ck_inference_calls_session_id",
+        ),
+        sa.CheckConstraint(
+            "user_id IS NULL OR (user_id <> '' AND user_id = btrim(user_id, E' \\t\\n\\r'))",
+            name="ck_inference_calls_user_id",
+        ),
+        sa.CheckConstraint(
+            "(input_tokens IS NULL OR input_tokens >= 0) AND (output_tokens IS NULL OR output_tokens >= 0) AND (duration_ms IS NULL OR duration_ms >= 0)",
+            name="ck_inference_calls_usage",
+        ),
+        sa.CheckConstraint(
+            "schema_version = 1", name="ck_inference_calls_schema_version"
+        ),
+        sa.PrimaryKeyConstraint("inference_call_id"),
+    )
+    op.create_index(
+        "ix_inference_calls_org_time",
+        "inference_calls",
+        ["org_id", "observed_at"],
+        unique=False,
+    )
+    op.create_index(
+        "uq_inference_calls_model_call",
+        "inference_calls",
+        ["org_id", "gateway_provider", "model_call_id"],
+        unique=True,
+        postgresql_where=sa.text("model_call_id IS NOT NULL"),
+    )
+    op.create_table(
+        "pushes",
+        sa.Column("push_id", sa.Text(), nullable=False),
+        sa.Column("org_id", sa.Text(), nullable=False),
+        sa.Column("provider", sa.Text(), nullable=False),
+        sa.Column("repo", sa.Text(), nullable=False),
+        sa.Column("clone_url", sa.Text(), nullable=False),
+        sa.Column("ref", sa.Text(), nullable=False),
+        sa.Column("before_sha", sa.Text(), nullable=False),
+        sa.Column("after_sha", sa.Text(), nullable=False),
+        sa.Column("forced", sa.Boolean(), nullable=False),
+        sa.Column("captured_at", sa.DateTime(timezone=True), nullable=False),
+        sa.CheckConstraint(
+            "after_sha ~ '^(?:[0-9a-f]{40}|[0-9a-f]{64})$'", name="ck_pushes_after_sha"
+        ),
+        sa.CheckConstraint(
+            "before_sha ~ '^(?:[0-9a-f]{40}|[0-9a-f]{64})$'",
+            name="ck_pushes_before_sha",
+        ),
+        sa.CheckConstraint(
+            "org_id ~ '^[a-z0-9][a-z0-9._-]{0,63}$'", name="ck_pushes_org_id"
+        ),
+        sa.CheckConstraint("provider IN ('github')", name="ck_pushes_provider"),
+        sa.CheckConstraint(
+            "push_id <> '' AND push_id = btrim(push_id, E' \\t\\n\\r')",
+            name="ck_pushes_push_id",
+        ),
+        sa.CheckConstraint(
+            "ref <> '' AND ref = btrim(ref, E' \\t\\n\\r')", name="ck_pushes_ref"
+        ),
+        sa.CheckConstraint(
+            "repo = '' OR (repo = lower(repo) AND repo ~ '^[^/]+/[^/]+$')",
+            name="ck_pushes_repo",
+        ),
+        sa.PrimaryKeyConstraint("push_id"),
+    )
+    op.create_index(
+        "uq_pushes_natural",
+        "pushes",
+        ["org_id", "repo", "ref", "before_sha", "after_sha"],
+        unique=True,
+    )
+    op.create_table(
+        "rejected_edits",
+        sa.Column("rejection_id", sa.Text(), nullable=False),
+        sa.Column("org_id", sa.Text(), nullable=False),
+        sa.Column("session_id", sa.Text(), nullable=False),
+        sa.Column("user_id", sa.Text(), nullable=True),
+        sa.Column("agent_harness", sa.Text(), nullable=False),
+        sa.Column("file_path", sa.Text(), nullable=False),
+        sa.Column("call_id", sa.Text(), nullable=False),
+        sa.Column("proposed", sa.Text(), nullable=False),
+        sa.Column("occurred_at", sa.DateTime(timezone=True), nullable=False),
+        sa.Column("captured_at", sa.DateTime(timezone=True), nullable=False),
+        sa.Column("raw", sa.Text(), nullable=False),
+        sa.CheckConstraint(
+            "agent_harness IN ('claude-code', 'copilot', 'codex', 'pi')",
+            name="ck_rejected_edits_agent_harness",
+        ),
+        sa.CheckConstraint(
+            "call_id <> '' AND call_id = btrim(call_id, E' \\t\\n\\r')",
+            name="ck_rejected_edits_call_id",
+        ),
+        sa.CheckConstraint(
+            "org_id ~ '^[a-z0-9][a-z0-9._-]{0,63}$'", name="ck_rejected_edits_org_id"
+        ),
+        sa.CheckConstraint(
+            "rejection_id <> '' AND rejection_id = btrim(rejection_id, E' \\t\\n\\r')",
+            name="ck_rejected_edits_rejection_id",
+        ),
+        sa.CheckConstraint(
+            "session_id <> '' AND session_id = btrim(session_id, E' \\t\\n\\r')",
+            name="ck_rejected_edits_session_id",
+        ),
+        sa.CheckConstraint(
+            "user_id IS NULL OR (user_id <> '' AND user_id = btrim(user_id, E' \\t\\n\\r'))",
+            name="ck_rejected_edits_user_id",
+        ),
+        sa.PrimaryKeyConstraint("rejection_id"),
+    )
+    op.create_index(
+        "uq_rejected_edits_call",
+        "rejected_edits",
+        ["org_id", "agent_harness", "session_id", "call_id"],
+        unique=True,
+    )
+    op.create_table(
+        "sessions",
+        sa.Column("org_id", sa.Text(), nullable=False),
+        sa.Column("session_id", sa.Text(), nullable=False),
+        sa.Column("user_id", sa.Text(), nullable=True),
+        sa.Column(
+            "user_id_conflict", sa.Boolean(), server_default="false", nullable=False
+        ),
+        sa.Column("first_observed_at", sa.DateTime(timezone=True), nullable=False),
+        sa.Column("last_observed_at", sa.DateTime(timezone=True), nullable=False),
+        sa.CheckConstraint(
+            "org_id ~ '^[a-z0-9][a-z0-9._-]{0,63}$'", name="ck_sessions_org_id"
+        ),
+        sa.CheckConstraint(
+            "session_id <> '' AND session_id = btrim(session_id, E' \\t\\n\\r')",
+            name="ck_sessions_session_id",
+        ),
+        sa.CheckConstraint(
+            "user_id IS NULL OR (user_id <> '' AND user_id = btrim(user_id, E' \\t\\n\\r'))",
+            name="ck_sessions_user_id",
+        ),
+        sa.CheckConstraint(
+            "first_observed_at <= last_observed_at",
+            name="ck_sessions_observation_bounds",
+        ),
+        sa.PrimaryKeyConstraint("org_id", "session_id", name="pk_sessions"),
+    )
+    # ### end Alembic commands ###
+
+
+def downgrade() -> None:
+    raise NotImplementedError("Sediment migrations are forward-only")