refactor: Refactor storage and new serve template (#947)

This commit is contained in:
Fangyin Cheng
2023-12-18 19:30:40 +08:00
committed by GitHub
parent 22d95b444b
commit 511a43b849
63 changed files with 1891 additions and 229 deletions

View File

@@ -1,5 +1,17 @@
from .utils import (
get_gpu_memory,
server_error_msg,
get_or_create_event_loop,
)
from .pagination_utils import PaginationResult
from .parameter_utils import BaseParameters, ParameterDescription, EnvArgumentParser
from .config_utils import AppConfig
__ALL__ = [
"get_gpu_memory",
"get_or_create_event_loop",
"PaginationResult",
"BaseParameters",
"ParameterDescription",
"EnvArgumentParser",
"AppConfig",
]

View File

@@ -51,19 +51,50 @@ def create_alembic_config(
def create_migration_script(
alembic_cfg: AlembicConfig, engine: Engine, message: str = "New migration"
) -> None:
alembic_cfg: AlembicConfig,
engine: Engine,
message: str = "New migration",
create_new_revision_if_noting_to_update: Optional[bool] = True,
) -> str:
"""Create migration script.
Args:
alembic_cfg: alembic config
engine: sqlalchemy engine
message: migration message
create_new_revision_if_noting_to_update: Whether to create a new revision if there is nothing to update,
pass False to avoid creating a new revision if there is nothing to update, default is True
Returns:
The path of the generated migration script.
"""
from alembic.script import ScriptDirectory
from alembic.runtime.migration import MigrationContext
# Check if the database is up-to-date
script_dir = ScriptDirectory.from_config(alembic_cfg)
with engine.connect() as connection:
alembic_cfg.attributes["connection"] = connection
command.revision(alembic_cfg, message, autogenerate=True)
context = MigrationContext.configure(connection=connection)
current_rev = context.get_current_revision()
head_rev = script_dir.get_current_head()
logger.info(
f"alembic migration current revision: {current_rev}, latest revision: {head_rev}"
)
should_create_revision = (
(current_rev is None and head_rev is None)
or current_rev != head_rev
or create_new_revision_if_noting_to_update
)
if should_create_revision:
with engine.connect() as connection:
alembic_cfg.attributes["connection"] = connection
revision = command.revision(alembic_cfg, message=message, autogenerate=True)
# Return the path of the generated migration script
return revision.path
elif current_rev == head_rev:
logger.info("No migration script to generate, database is up-to-date")
# If no new revision is created, return None or an appropriate message
return None
def upgrade_database(
@@ -82,6 +113,37 @@ def upgrade_database(
command.upgrade(alembic_cfg, target_version)
def generate_sql_for_upgrade(
alembic_cfg: AlembicConfig,
engine: Engine,
target_version: Optional[str] = "head",
output_file: Optional[str] = "migration.sql",
) -> None:
"""Generate SQL for upgrading database to target version.
Args:
alembic_cfg: alembic config
engine: sqlalchemy engine
target_version: target version, default is head (latest version)
output_file: file to write the SQL script
TODO: Can't generate SQL for most of the operations.
"""
import contextlib
import io
with engine.connect() as connection, contextlib.redirect_stdout(
io.StringIO()
) as stdout:
alembic_cfg.attributes["connection"] = connection
# Generating SQL instead of applying changes
command.upgrade(alembic_cfg, target_version, sql=True)
# Write the generated SQL to a file
with open(output_file, "w", encoding="utf-8") as file:
file.write(stdout.getvalue())
def downgrade_database(
alembic_cfg: AlembicConfig, engine: Engine, revision: str = "-1"
):
@@ -160,9 +222,94 @@ or
rm -rf pilot/meta_data/alembic/versions/*
rm -rf pilot/meta_data/alembic/dbgpt.db
```
If your database is a shared database, and you run DB-GPT in multiple instances,
you should make sure that all migration scripts are same in all instances, in this case,
wo strongly recommend you close migration feature by setting `--disable_alembic_upgrade`.
and use `dbgpt db migration` command to manage migration scripts.
"""
def _check_database_migration_status(alembic_cfg: AlembicConfig, engine: Engine):
"""Check if the database is at the latest migration revision.
If your database is a shared database, and you run DB-GPT in multiple instances,
you should make sure that all migration scripts are same in all instances, in this case,
wo strongly recommend you close migration feature by setting `disable_alembic_upgrade` to True.
and use `dbgpt db migration` command to manage migration scripts.
Args:
alembic_cfg: Alembic configuration object.
engine: SQLAlchemy engine instance.
Raises:
Exception: If the database is not at the latest revision.
"""
from alembic.script import ScriptDirectory
from alembic.runtime.migration import MigrationContext
script = ScriptDirectory.from_config(alembic_cfg)
def get_current_revision(engine):
with engine.connect() as connection:
context = MigrationContext.configure(connection=connection)
return context.get_current_revision()
current_rev = get_current_revision(engine)
head_rev = script.get_current_head()
script_info_msg = "Migration versions and their file paths:"
script_info_msg += f"\n{'='*40}Migration versions{'='*40}\n"
for revision in script.walk_revisions(base="base"):
current_marker = "(current)" if revision.revision == current_rev else ""
script_path = script.get_revision(revision.revision).path
script_info_msg += f"\n{revision.revision} {current_marker}: {revision.doc} (Path: {script_path})"
script_info_msg += f"\n{'='*90}"
logger.info(script_info_msg)
if current_rev != head_rev:
logger.error(
"Database is not at the latest revision. "
f"Current revision: {current_rev}, latest revision: {head_rev}\n"
"Please apply existing migration scripts before generating new ones. "
"Check the listed file paths for migration scripts.\n"
f"Also you can try the following solutions:\n{_MIGRATION_SOLUTION}\n"
)
raise Exception(
"Check database migration status failed, you can see the error and solutions above"
)
def _get_latest_revision(alembic_cfg: AlembicConfig, engine: Engine) -> str:
"""Get the latest revision of the database.
Args:
alembic_cfg: Alembic configuration object.
engine: SQLAlchemy engine instance.
Returns:
The latest revision as a string.
"""
from alembic.runtime.migration import MigrationContext
with engine.connect() as connection:
context = MigrationContext.configure(connection=connection)
return context.get_current_revision()
def _delete_migration_script(script_path: str):
"""Delete a migration script.
Args:
script_path: The path of the migration script to delete.
"""
if os.path.exists(script_path):
os.remove(script_path)
logger.info(f"Deleted migration script at: {script_path}")
else:
logger.warning(f"Migration script not found at: {script_path}")
def _ddl_init_and_upgrade(
default_meta_data_path: str,
disable_alembic_upgrade: bool,
@@ -203,7 +350,19 @@ def _ddl_init_and_upgrade(
script_location,
)
try:
create_migration_script(alembic_cfg, db.engine)
_check_database_migration_status(alembic_cfg, db.engine)
except Exception as e:
logger.error(f"Failed to check database migration status: {e}")
raise
latest_revision_before = "__latest_revision_before__"
new_script_path = None
try:
latest_revision_before = _get_latest_revision(alembic_cfg, db.engine)
# create_new_revision_if_noting_to_update=False avoid creating a lot of empty migration scripts
# TODO Set create_new_revision_if_noting_to_update=False, not working now.
new_script_path = create_migration_script(
alembic_cfg, db.engine, create_new_revision_if_noting_to_update=True
)
upgrade_database(alembic_cfg, db.engine)
except CommandError as e:
if "Target database is not up to date" in str(e):
@@ -216,4 +375,10 @@ def _ddl_init_and_upgrade(
"you can see the error and solutions above"
) from e
else:
latest_revision_after = _get_latest_revision(alembic_cfg, db.engine)
if latest_revision_before != latest_revision_after:
logger.error(
f"Upgrade database failed. Please review the migration script manually. "
f"Failed script path: {new_script_path}\nError: {e}"
)
raise e

View File

@@ -0,0 +1,32 @@
from functools import cache
from typing import Any, Dict, Optional
class AppConfig:
def __init__(self):
self.configs = {}
def set(self, key: str, value: Any) -> None:
"""Set config value by key
Args:
key (str): The key of config
value (Any): The value of config
"""
self.configs[key] = value
def get(self, key, default: Optional[Any] = None) -> Any:
"""Get config value by key
Args:
key (str): The key of config
default (Optional[Any], optional): The default value if key not found. Defaults to None.
"""
return self.configs.get(key, default)
@cache
def get_all_by_prefix(self, prefix) -> Dict[str, Any]:
"""Get all config values by prefix
Args:
prefix (str): The prefix of config
"""
return {k: v for k, v in self.configs.items() if k.startswith(prefix)}

View File

@@ -6,7 +6,6 @@ import logging.handlers
from typing import Any, List
import os
import sys
import asyncio
from dbgpt.configs.model_config import LOGDIR
@@ -81,17 +80,6 @@ def _build_logger(logger_name, logging_level=None, logger_filename: str = None):
setup_logging_level(logging_level=logging_level)
logging.getLogger().handlers[0].setFormatter(formatter)
# Redirect stdout and stderr to loggers
# stdout_logger = logging.getLogger("stdout")
# stdout_logger.setLevel(logging.INFO)
# sl_1 = StreamToLogger(stdout_logger, logging.INFO)
# sys.stdout = sl_1
#
# stderr_logger = logging.getLogger("stderr")
# stderr_logger.setLevel(logging.ERROR)
# sl = StreamToLogger(stderr_logger, logging.ERROR)
# sys.stderr = sl
# Add a file handler for all loggers
if handler is None and logger_filename:
os.makedirs(LOGDIR, exist_ok=True)