Configuration API

Pydantic settings and datasource configuration.

class sqldbagent.core.config.PoolSettings(**data)[source]

Bases: BaseModel

Database pool settings.

Variables:

  • size – Base pool size for sync engine usage.

  • max_overflow – Additional overflow connections allowed above size.

  • timeout_seconds – Time to wait for a pooled connection.

Parameters:
model_config: ClassVar[ConfigDict] = {'extra': 'forbid'}

Configuration for the model, should be a dictionary conforming to [ConfigDict][pydantic.config.ConfigDict].

size: int
max_overflow: int
timeout_seconds: float
class sqldbagent.core.config.SafetySettings(**data)[source]

Bases: BaseModel

Default SQL safety policy settings.

Variables:

  • read_only – Whether execution paths should default to read-only behavior.

  • allow_writes – Whether writable execution may be requested explicitly on top of the read-only default.

  • statement_timeout_seconds – Default execution timeout.

  • max_rows – Default maximum row count for user-facing query surfaces.

  • allowed_schemas – Optional schema allowlist enforced by higher layers.

Parameters:
model_config: ClassVar[ConfigDict] = {'extra': 'forbid'}

Configuration for the model, should be a dictionary conforming to [ConfigDict][pydantic.config.ConfigDict].

read_only: bool
allow_writes: bool
statement_timeout_seconds: float
max_rows: int
allowed_schemas: list[str]
class sqldbagent.core.config.ProfilingSettings(**data)[source]

Bases: BaseModel

Default profiling settings.

Variables:

  • default_sample_size – Default number of sample rows.

  • max_sample_size – Maximum sample rows allowed by default tooling.

  • exact_unique_counts – Whether generic profilers should compute exact unique counts.

Parameters:

  • data (Any)

  • default_sample_size (int)

  • max_sample_size (int)

  • exact_unique_counts (bool)

model_config: ClassVar[ConfigDict] = {'extra': 'forbid'}

Configuration for the model, should be a dictionary conforming to [ConfigDict][pydantic.config.ConfigDict].

default_sample_size: int
max_sample_size: int
exact_unique_counts: bool
class sqldbagent.core.config.ArtifactSettings(**data)[source]

Bases: BaseModel

Artifact persistence settings.

Variables:

  • root_dir – Base directory for generated artifacts.

  • snapshots_dir – Snapshot subdirectory under root_dir.

  • documents_dir – Document-export subdirectory under root_dir.

  • diagrams_dir – Diagram-export subdirectory under root_dir.

  • prompts_dir – Prompt-export subdirectory under root_dir.

  • prompt_enhancements_dir – Prompt-enhancement subdirectory under root_dir.

  • embeddings_cache_dir – Embedding cache subdirectory under root_dir.

  • vectorstores_dir – Retrieval/vectorstore manifest subdirectory under root_dir.

Parameters:

  • data (Any)

  • root_dir (str)

  • snapshots_dir (str)

  • documents_dir (str)

  • diagrams_dir (str)

  • prompts_dir (str)

  • prompt_enhancements_dir (str)

  • embeddings_cache_dir (str)

  • vectorstores_dir (str)

model_config: ClassVar[ConfigDict] = {'extra': 'forbid'}

Configuration for the model, should be a dictionary conforming to [ConfigDict][pydantic.config.ConfigDict].

root_dir: str
snapshots_dir: str
documents_dir: str
diagrams_dir: str
prompts_dir: str
prompt_enhancements_dir: str
embeddings_cache_dir: str
vectorstores_dir: str
class sqldbagent.core.config.EmbeddingSettings(**data)[source]

Bases: BaseModel

Embedding-provider settings.

Variables:

  • provider – Embedding backend identifier.

  • model – Provider-specific embedding model name.

  • dimensions – Optional embedding dimensionality override.

  • batch_size – Batch size used by embedding providers when supported.

  • cache_query_embeddings – Whether query embeddings should also be cached.

Parameters:

  • data (Any)

  • provider (Literal['openai', 'hash'])

  • model (str)

  • dimensions (int | None)

  • batch_size (int)

  • cache_query_embeddings (bool)

model_config: ClassVar[ConfigDict] = {'extra': 'ignore', 'populate_by_name': True, 'validate_by_alias': True, 'validate_by_name': True}

Configuration for the model, should be a dictionary conforming to [ConfigDict][pydantic.config.ConfigDict].

provider: Literal['openai', 'hash']
model: str
dimensions: int | None
batch_size: int
cache_query_embeddings: bool
class sqldbagent.core.config.RetrievalSettings(**data)[source]

Bases: BaseModel

Retrieval and vectorstore settings.

Variables:

  • backend – Retrieval backend identifier.

  • qdrant_url – Base Qdrant HTTP URL.

  • qdrant_api_key – Optional Qdrant API key.

  • qdrant_grpc_port – Qdrant gRPC port when gRPC transport is enabled.

  • qdrant_prefer_grpc – Whether to prefer gRPC transport.

  • collection_prefix – Prefix used when generating Qdrant collection names.

  • default_top_k – Default number of documents returned from retrieval.

  • default_fetch_k – Fetch pool size used by MMR retrieval.

  • use_mmr – Whether retrievers should prefer MMR over plain similarity.

  • score_threshold – Optional score threshold applied to retrieval.

  • create_payload_indexes – Whether payload indexes should be created in Qdrant.

Parameters:

  • data (Any)

  • backend (Literal['qdrant'])

  • qdrant_url (str)

  • qdrant_api_key (str | None)

  • qdrant_grpc_port (int)

  • qdrant_prefer_grpc (bool)

  • collection_prefix (str)

  • default_top_k (int)

  • default_fetch_k (int)

  • use_mmr (bool)

  • score_threshold (float | None)

  • create_payload_indexes (bool)

model_config: ClassVar[ConfigDict] = {'extra': 'ignore', 'populate_by_name': True, 'validate_by_alias': True, 'validate_by_name': True}

Configuration for the model, should be a dictionary conforming to [ConfigDict][pydantic.config.ConfigDict].

backend: Literal['qdrant']
qdrant_url: str
qdrant_api_key: str | None
qdrant_grpc_port: int
qdrant_prefer_grpc: bool
collection_prefix: str
default_top_k: int
default_fetch_k: int
use_mmr: bool
score_threshold: float | None
create_payload_indexes: bool
class sqldbagent.core.config.LLMSettings(**data)[source]

Bases: BaseModel

Optional model-provider settings.

Variables:

  • default_provider – Default provider identifier for future LLM features.

  • default_model – Default model name for future LLM features.

  • reasoning_effort – Default reasoning effort for supported reasoning models.

  • openai_api_key – Optional OpenAI API key loaded from the environment.

  • openai_base_url – Optional OpenAI-compatible base URL.

  • anthropic_api_key – Optional Anthropic API key loaded from the environment.

  • anthropic_base_url – Optional Anthropic base URL override.

Parameters:

  • data (Any)

  • default_provider (str | None)

  • default_model (str | None)

  • reasoning_effort (str | None)

  • openai_api_key (str | None)

  • openai_base_url (str | None)

  • anthropic_api_key (str | None)

  • anthropic_base_url (str | None)

model_config: ClassVar[ConfigDict] = {'extra': 'ignore', 'populate_by_name': True, 'validate_by_alias': True, 'validate_by_name': True}

Configuration for the model, should be a dictionary conforming to [ConfigDict][pydantic.config.ConfigDict].

default_provider: str | None
default_model: str | None
reasoning_effort: str | None
openai_api_key: str | None
openai_base_url: str | None
anthropic_api_key: str | None
anthropic_base_url: str | None
class sqldbagent.core.config.LangSmithSettings(**data)[source]

Bases: BaseModel

LangSmith tracing and project settings.

Variables:

  • tracing – Whether LangSmith tracing is enabled for supported surfaces.

  • project – LangSmith project name used for traces.

  • api_key – Optional LangSmith API key loaded from the environment.

  • endpoint – Optional LangSmith API endpoint override.

  • workspace_id – Optional LangSmith workspace identifier.

  • tags – Default LangSmith trace tags applied across surfaces.

Parameters:
model_config: ClassVar[ConfigDict] = {'extra': 'ignore', 'populate_by_name': True, 'validate_by_alias': True, 'validate_by_name': True}

Configuration for the model, should be a dictionary conforming to [ConfigDict][pydantic.config.ConfigDict].

tracing: bool
project: str
api_key: str | None
endpoint: str | None
workspace_id: str | None
tags: list[str]
classmethod validate_tags(value)[source]

Normalize tag values from env strings or iterables.

Parameters:

value (object) – Raw LangSmith tags value.

Returns:

Normalized list value for Pydantic parsing.

Return type:

object

class sqldbagent.core.config.AgentCheckpointSettings(**data)[source]

Bases: BaseModel

Agent checkpoint persistence settings.

Variables:

  • backend – Checkpoint backend to use for agent persistence.

  • postgres_url – Optional Postgres connection string for LangGraph checkpointing.

  • auto_setup – Whether Postgres checkpointer tables should be initialized automatically.

  • pipeline – Whether the Postgres saver should use pipelining when supported.

Parameters:

  • data (Any)

  • backend (Literal['memory', 'postgres'])

  • postgres_url (str | None)

  • auto_setup (bool)

  • pipeline (bool)

model_config: ClassVar[ConfigDict] = {'extra': 'ignore', 'populate_by_name': True, 'validate_by_alias': True, 'validate_by_name': True}

Configuration for the model, should be a dictionary conforming to [ConfigDict][pydantic.config.ConfigDict].

backend: Literal['memory', 'postgres']
postgres_url: str | None
auto_setup: bool
pipeline: bool
class sqldbagent.core.config.AgentMemorySettings(**data)[source]

Bases: BaseModel

Agent long-term memory store settings.

Variables:

  • backend – Memory-store backend used by LangGraph runtime.store.

  • postgres_url – Optional Postgres connection string for durable memory.

  • auto_setup – Whether Postgres store tables should be initialized automatically.

  • pipeline – Whether the Postgres store should use pipelining when supported.

  • namespace_prefix – Prefix used when building sqldbagent store namespaces.

  • auto_sync_from_snapshot – Whether agents should persist snapshot-derived datasource context into the long-term store automatically.

  • auto_create_snapshot_if_missing – Whether state-seeding should create a snapshot for the active schema when no stored snapshot exists yet.

  • include_in_system_prompt – Whether remembered store context should be merged into the effective system prompt.

Parameters:

  • data (Any)

  • backend (Literal['disabled', 'memory', 'postgres'])

  • postgres_url (str | None)

  • auto_setup (bool)

  • pipeline (bool)

  • namespace_prefix (str)

  • auto_sync_from_snapshot (bool)

  • auto_create_snapshot_if_missing (bool)

  • include_in_system_prompt (bool)

model_config: ClassVar[ConfigDict] = {'extra': 'ignore', 'populate_by_name': True, 'validate_by_alias': True, 'validate_by_name': True}

Configuration for the model, should be a dictionary conforming to [ConfigDict][pydantic.config.ConfigDict].

backend: Literal['disabled', 'memory', 'postgres']
postgres_url: str | None
auto_setup: bool
pipeline: bool
namespace_prefix: str
auto_sync_from_snapshot: bool
auto_create_snapshot_if_missing: bool
include_in_system_prompt: bool
class sqldbagent.core.config.AgentSettings(**data)[source]

Bases: BaseModel

Agent orchestration settings.

Variables:

  • name – Stable agent name for LangChain/LangGraph surfaces.

  • include_latest_snapshot_context – Whether agents should inject latest snapshot summaries.

  • max_model_calls_per_run – Optional cap for model calls in a single run.

  • max_tool_calls_per_run – Optional cap for tool calls in a single run.

  • enable_todo_middleware – Whether the LangChain todo middleware should be enabled.

  • enable_human_in_the_loop – Whether safe_query_sql should require approval middleware.

  • enable_summarization_middleware – Whether context summarization middleware is enabled.

  • summarization_trigger_fraction – Fractional context threshold for summarization.

  • summarization_keep_messages – Number of recent messages to preserve after summarization.

  • summarization_model – Optional dedicated model identifier for summarization.

  • tool_call_digest_limit – Maximum number of compressed tool-call summaries to retain.

  • checkpoint – Agent checkpoint persistence settings.

  • memory – Agent long-term memory store settings.

  • enable_prompt_enhancements – Whether dynamic prompts should merge saved prompt-enhancement artifacts.

Parameters:

  • data (Any)

  • name (str)

  • include_latest_snapshot_context (bool)

  • max_model_calls_per_run (int | None)

  • max_tool_calls_per_run (int | None)

  • enable_todo_middleware (bool)

  • enable_human_in_the_loop (bool)

  • enable_summarization_middleware (bool)

  • summarization_trigger_fraction (float)

  • summarization_keep_messages (int)

  • summarization_model (str | None)

  • tool_call_digest_limit (int)

  • checkpoint (AgentCheckpointSettings)

  • memory (AgentMemorySettings)

  • enable_prompt_enhancements (bool)

model_config: ClassVar[ConfigDict] = {'extra': 'ignore', 'populate_by_name': True, 'validate_by_alias': True, 'validate_by_name': True}

Configuration for the model, should be a dictionary conforming to [ConfigDict][pydantic.config.ConfigDict].

name: str
include_latest_snapshot_context: bool
max_model_calls_per_run: int | None
max_tool_calls_per_run: int | None
enable_todo_middleware: bool
enable_human_in_the_loop: bool
enable_summarization_middleware: bool
summarization_trigger_fraction: float
summarization_keep_messages: int
summarization_model: str | None
tool_call_digest_limit: int
checkpoint: AgentCheckpointSettings
memory: AgentMemorySettings
enable_prompt_enhancements: bool
class sqldbagent.core.config.MCPSettings(**data)[source]

Bases: BaseModel

FastMCP server settings.

Variables:

  • transport – Default MCP transport to serve.

  • host – Default host for HTTP-based transports.

  • port – Default port for HTTP-based transports.

  • path – Default HTTP path for streamable transports.

  • log_level – Default FastMCP/Uvicorn log level for HTTP transports.

  • show_banner – Whether the FastMCP banner should be shown on startup.

  • stateless_http – Whether streamable HTTP should run in stateless mode.

Parameters:

  • data (Any)

  • transport (Literal['stdio', 'http', 'sse', 'streamable-http'])

  • host (str)

  • port (int)

  • path (str)

  • log_level (str)

  • show_banner (bool)

  • stateless_http (bool)

model_config: ClassVar[ConfigDict] = {'extra': 'ignore', 'populate_by_name': True, 'validate_by_alias': True, 'validate_by_name': True}

Configuration for the model, should be a dictionary conforming to [ConfigDict][pydantic.config.ConfigDict].

transport: Literal['stdio', 'http', 'sse', 'streamable-http']
host: str
port: int
path: str
log_level: str
show_banner: bool
stateless_http: bool
class sqldbagent.core.config.DatasourceSettings(**data)[source]

Bases: BaseModel

Single datasource definition.

Variables:

  • name – Stable datasource identifier used by services and adapters.

  • dialect – Database dialect used by the datasource.

  • url – SQLAlchemy-compatible connection URL.

  • echo – Whether SQLAlchemy should emit SQL logs.

  • pool – Pool configuration for this datasource.

  • safety – Default safety policy for this datasource.

Parameters:
model_config: ClassVar[ConfigDict] = {'extra': 'forbid'}

Configuration for the model, should be a dictionary conforming to [ConfigDict][pydantic.config.ConfigDict].

name: str
dialect: Dialect
url: str
echo: bool
pool: PoolSettings
safety: SafetySettings
classmethod validate_name(value)[source]

Validate and normalize datasource names.

Parameters:

value (str) – Raw datasource name.

Returns:

Normalized datasource name.

Return type:

str

class sqldbagent.core.config.AppSettings(_case_sensitive=None, _nested_model_default_partial_update=None, _env_prefix=None, _env_prefix_target=None, _env_file=PosixPath('.'), _env_file_encoding=None, _env_ignore_empty=None, _env_nested_delimiter=None, _env_nested_max_split=None, _env_parse_none_str=None, _env_parse_enums=None, _cli_prog_name=None, _cli_parse_args=None, _cli_settings_source=None, _cli_parse_none_str=None, _cli_hide_none_type=None, _cli_avoid_json=None, _cli_enforce_required=None, _cli_use_class_docs_for_groups=None, _cli_exit_on_error=None, _cli_prefix=None, _cli_flag_prefix_char=None, _cli_implicit_flags=None, _cli_ignore_unknown_args=None, _cli_kebab_case=None, _cli_shortcuts=None, _secrets_dir=None, _build_sources=None, **values)[source]

Bases: BaseSettings

Top-level application settings.

Variables:

  • env – Deployment environment label.

  • log_level – Default application log level.

  • datasources – Configured datasource definitions.

  • datasource_aliases – Optional alias map from short names to datasource names.

  • profiling – Default profiling settings.

  • artifacts – Artifact persistence settings.

  • llm – Optional model-provider settings.

  • langsmith – LangSmith tracing settings.

  • embeddings – Embedding-provider settings.

  • retrieval – Retrieval/vectorstore settings.

  • agent – Agent orchestration settings.

  • mcp – FastMCP server settings.

Parameters:

  • _case_sensitive (bool | None, default: None)

  • _nested_model_default_partial_update (bool | None, default: None)

  • _env_prefix (str | None, default: None)

  • _env_prefix_target (Literal['variable', 'alias', 'all'] | None, default: None)

  • _env_file (Path | str | Sequence[Path | str] | None, default: PosixPath('.'))

  • _env_file_encoding (str | None, default: None)

  • _env_ignore_empty (bool | None, default: None)

  • _env_nested_delimiter (str | None, default: None)

  • _env_nested_max_split (int | None, default: None)

  • _env_parse_none_str (str | None, default: None)

  • _env_parse_enums (bool | None, default: None)

  • _cli_prog_name (str | None, default: None)

  • _cli_parse_args (bool | list[str] | tuple[str, ...] | None, default: None)

  • _cli_settings_source (CliSettingsSource[Any] | None, default: None)

  • _cli_parse_none_str (str | None, default: None)

  • _cli_hide_none_type (bool | None, default: None)

  • _cli_avoid_json (bool | None, default: None)

  • _cli_enforce_required (bool | None, default: None)

  • _cli_use_class_docs_for_groups (bool | None, default: None)

  • _cli_exit_on_error (bool | None, default: None)

  • _cli_prefix (str | None, default: None)

  • _cli_flag_prefix_char (str | None, default: None)

  • _cli_implicit_flags (bool | Literal['dual', 'toggle'] | None, default: None)

  • _cli_ignore_unknown_args (bool | None, default: None)

  • _cli_kebab_case (bool | Literal['all', 'no_enums'] | None, default: None)

  • _cli_shortcuts (Mapping[str, str | list[str]] | None, default: None)

  • _secrets_dir (Path | str | Sequence[Path | str] | None, default: None)

  • _build_sources (tuple[tuple[PydanticBaseSettingsSource, ...], dict[str, Any]] | None, default: None)

  • values (Any)

  • env (str)

  • log_level (str)

  • datasources (list[DatasourceSettings])

  • datasource_aliases (dict[str, str])

  • profiling (ProfilingSettings)

  • artifacts (ArtifactSettings)

  • llm (LLMSettings)

  • langsmith (LangSmithSettings)

  • embeddings (EmbeddingSettings)

  • retrieval (RetrievalSettings)

  • agent (AgentSettings)

  • mcp (MCPSettings)

  • default_datasource_name (str | None)

  • default_schema_name (str | None)

  • llm_default_provider (str | None)

  • llm_default_model (str | None)

  • llm_reasoning_effort (str | None)

  • openai_api_key (str | None)

  • openai_base_url (str | None)

  • anthropic_api_key (str | None)

  • anthropic_base_url (str | None)

  • langsmith_tracing (bool | None)

  • langsmith_project (str | None)

  • langsmith_api_key (str | None)

  • langsmith_endpoint (str | None)

  • langsmith_workspace_id (str | None)

  • langsmith_tags (list[str] | str | None)

  • agent_name (str | None)

  • agent_include_latest_snapshot_context (bool | None)

  • agent_max_model_calls_per_run (int | None)

  • agent_max_tool_calls_per_run (int | None)

  • agent_enable_todo_middleware (bool | None)

  • agent_enable_human_in_the_loop (bool | None)

  • agent_enable_summarization_middleware (bool | None)

  • agent_summarization_trigger_fraction (float | None)

  • agent_summarization_keep_messages (int | None)

  • agent_summarization_model (str | None)

  • agent_tool_call_digest_limit (int | None)

  • agent_checkpoint_backend (Literal['memory', 'postgres'] | None)

  • agent_checkpoint_postgres_url (str | None)

  • agent_checkpoint_auto_setup (bool | None)

  • agent_checkpoint_pipeline (bool | None)

  • agent_memory_backend (Literal['disabled', 'memory', 'postgres'] | None)

  • agent_memory_postgres_url (str | None)

  • agent_memory_auto_setup (bool | None)

  • agent_memory_pipeline (bool | None)

  • agent_memory_namespace_prefix (str | None)

  • agent_memory_auto_sync_from_snapshot (bool | None)

  • agent_memory_auto_create_snapshot_if_missing (bool | None)

  • agent_memory_include_in_system_prompt (bool | None)

  • qdrant_url (str | None)

  • qdrant_api_key (str | None)

  • qdrant_grpc_port (int | None)

  • qdrant_prefer_grpc (bool | None)

  • embeddings_provider (Literal['openai', 'hash'] | None)

  • embeddings_model (str | None)

  • embeddings_dimensions (int | None)

  • embeddings_batch_size (int | None)

  • embeddings_cache_query_embeddings (bool | None)

  • sqlite_path (str | None)

  • postgres_host (str | None)

  • postgres_port (int)

  • postgres_db (str | None)

  • postgres_user (str | None)

  • postgres_password (str | None)

  • postgres_demo_host (str | None)

  • postgres_demo_port (int)

  • postgres_demo_db (str | None)

  • postgres_demo_user (str | None)

  • postgres_demo_password (str | None)

  • mssql_host (str | None)

  • mssql_port (int)

  • mssql_database (str | None)

  • mssql_user (str | None)

  • mssql_password (str | None)

  • mssql_driver (str)

model_config: ClassVar[SettingsConfigDict] = {'arbitrary_types_allowed': True, 'case_sensitive': False, 'cli_avoid_json': False, 'cli_enforce_required': False, 'cli_exit_on_error': True, 'cli_flag_prefix_char': '-', 'cli_hide_none_type': False, 'cli_ignore_unknown_args': False, 'cli_implicit_flags': False, 'cli_kebab_case': False, 'cli_parse_args': None, 'cli_parse_none_str': None, 'cli_prefix': '', 'cli_prog_name': None, 'cli_shortcuts': None, 'cli_use_class_docs_for_groups': False, 'enable_decoding': True, 'env_file': '.env', 'env_file_encoding': None, 'env_ignore_empty': False, 'env_nested_delimiter': '__', 'env_nested_max_split': None, 'env_parse_enums': None, 'env_parse_none_str': None, 'env_prefix': 'SQLDBAGENT_', 'env_prefix_target': 'variable', 'extra': 'ignore', 'json_file': None, 'json_file_encoding': None, 'nested_model_default_partial_update': False, 'populate_by_name': True, 'protected_namespaces': ('model_validate', 'model_dump', 'settings_customise_sources'), 'secrets_dir': None, 'toml_file': None, 'validate_by_alias': True, 'validate_by_name': True, 'validate_default': True, 'yaml_config_section': None, 'yaml_file': None, 'yaml_file_encoding': None}

Configuration for the model, should be a dictionary conforming to [ConfigDict][pydantic.config.ConfigDict].

env: str
log_level: str
datasources: list[DatasourceSettings]
datasource_aliases: dict[str, str]
profiling: ProfilingSettings
artifacts: ArtifactSettings
llm: LLMSettings
langsmith: LangSmithSettings
embeddings: EmbeddingSettings
retrieval: RetrievalSettings
agent: AgentSettings
mcp: MCPSettings
default_datasource_name: str | None
default_schema_name: str | None
llm_default_provider: str | None
llm_default_model: str | None
llm_reasoning_effort: str | None
openai_api_key: str | None
openai_base_url: str | None
anthropic_api_key: str | None
anthropic_base_url: str | None
langsmith_tracing: bool | None
langsmith_project: str | None
langsmith_api_key: str | None
langsmith_endpoint: str | None
langsmith_workspace_id: str | None
langsmith_tags: list[str] | str | None
agent_name: str | None
agent_include_latest_snapshot_context: bool | None
agent_max_model_calls_per_run: int | None
agent_max_tool_calls_per_run: int | None
agent_enable_todo_middleware: bool | None
agent_enable_human_in_the_loop: bool | None
agent_enable_summarization_middleware: bool | None
agent_summarization_trigger_fraction: float | None
agent_summarization_keep_messages: int | None
agent_summarization_model: str | None
agent_tool_call_digest_limit: int | None
agent_checkpoint_backend: Literal['memory', 'postgres'] | None
agent_checkpoint_postgres_url: str | None
agent_checkpoint_auto_setup: bool | None
agent_checkpoint_pipeline: bool | None
agent_memory_backend: Literal['disabled', 'memory', 'postgres'] | None
agent_memory_postgres_url: str | None
agent_memory_auto_setup: bool | None
agent_memory_pipeline: bool | None
agent_memory_namespace_prefix: str | None
agent_memory_auto_sync_from_snapshot: bool | None
agent_memory_auto_create_snapshot_if_missing: bool | None
agent_memory_include_in_system_prompt: bool | None
qdrant_url: str | None
qdrant_api_key: str | None
qdrant_grpc_port: int | None
qdrant_prefer_grpc: bool | None
embeddings_provider: Literal['openai', 'hash'] | None
embeddings_model: str | None
embeddings_dimensions: int | None
embeddings_batch_size: int | None
embeddings_cache_query_embeddings: bool | None
sqlite_path: str | None
postgres_host: str | None
postgres_port: int
postgres_db: str | None
postgres_user: str | None
postgres_password: str | None
postgres_demo_host: str | None
postgres_demo_port: int
postgres_demo_db: str | None
postgres_demo_user: str | None
postgres_demo_password: str | None
mssql_host: str | None
mssql_port: int
mssql_database: str | None
mssql_user: str | None
mssql_password: str | None
mssql_driver: str
classmethod validate_unique_datasources(value)[source]

Reject duplicate datasource names.

Parameters:

value (list[DatasourceSettings]) – Configured datasource definitions.

Returns:

Validated datasource definitions.

Return type:

list[DatasourceSettings]

classmethod validate_datasource_aliases(value)[source]

Validate datasource alias keys and targets.

Parameters:

value (dict[str, str]) – Alias mapping loaded from settings.

Returns:

Normalized alias mapping.

Return type:

dict[str, str]

build_default_datasources()[source]

Build convenience datasource definitions from environment fields.

Returns:

Settings with synthesized datasources when needed.

Return type:

AppSettings

resolve_datasource_name(name)[source]

Resolve a datasource name or alias to its canonical datasource name.

Parameters:

name (str) – Datasource name or alias.

Returns:

Canonical datasource name.

Return type:

str

get_datasource(name)[source]

Return a datasource by name.

Parameters:

name (str) – Datasource identifier.

Returns:

Matching datasource configuration.

Return type:

DatasourceSettings

Raises:

ConfigurationError – If the datasource is unknown.

resolve_default_datasource_name()[source]

Return the default datasource name for runtime surfaces.

Returns:

Preferred datasource name.

Return type:

str

Raises:

ConfigurationError – If no datasources are configured.

sqldbagent.core.config.load_settings()[source]

Load and cache application settings.

Returns:

Cached application settings.

Return type:

AppSettings