mirror/OpenLLM
1
0
Fork 0
mirror of https://github.com/bentoml/OpenLLM.git synced 2026-06-12 02:20:32 -04:00
Code Issues Packages Projects Releases Wiki Activity

fix: do not reply on env var for built bento/docker (#477)

Browse Source 
Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>
This commit is contained in:
Zhao Shenyang
2023-10-11 00:29:20 +08:00
committed by GitHub
parent d8da2cc3a5
commit bf96570eab
5 changed files with 38 additions and 17 deletions
Download Patch File Download Diff File Expand all files Collapse all files

19
openllm-python/src/openllm/bundle/_package.py
View File

@@ -124,10 +124,7 @@ def construct_docker_options(llm: openllm.LLM[t.Any, t.Any], _: FS, workers_per_
env_dict = {
env.backend: env['backend_value'],
env.config: f"'{llm.config.model_dump_json().decode()}'",
env.model_id: f'/home/bentoml/bento/models/{llm.tag.path()}',
'OPENLLM_MODEL': llm.config['model_name'],
'OPENLLM_SERIALIZATION': serialisation,
'OPENLLM_ADAPTER_MAP': f"'{orjson.dumps(adapter_map).decode()}'",
'BENTOML_DEBUG': str(True),
'BENTOML_QUIET': str(False),
'BENTOML_CONFIG_OPTIONS': f"'{environ['BENTOML_CONFIG_OPTIONS']}'",
@@ -143,6 +140,7 @@ def construct_docker_options(llm: openllm.LLM[t.Any, t.Any], _: FS, workers_per_
return DockerOptions(base_image=f'{oci.CONTAINER_NAMES[container_registry]}:{oci.get_base_container_tag(container_version_strategy)}', env=env_dict, dockerfile_template=dockerfile_template)
OPENLLM_MODEL_NAME = '# openllm: model name'
OPENLLM_MODEL_ID = '# openllm: model id'
OPENLLM_MODEL_ADAPTER_MAP = '# openllm: model adapter map'
class ModelNameFormatter(string.Formatter):
@@ -170,21 +168,30 @@ class ModelAdapterMapFormatter(ModelNameFormatter):
model_keyword: LiteralString = '__model_adapter_map__'
_service_file = Path(os.path.abspath(__file__)).parent.parent / '_service.py'
_service_vars_file = Path(os.path.abspath(__file__)).parent.parent / '_service_vars_pkg.py'
def write_service(llm: openllm.LLM[t.Any, t.Any], adapter_map: dict[str, str | None] | None, llm_fs: FS) -> None:
from openllm_core.utils import DEBUG
model_name = llm.config['model_name']
logger.debug('Generating service file for %s at %s (dir=%s)', model_name, llm.config['service_name'], llm_fs.getsyspath('/'))
with open(_service_file.__fspath__(), 'r') as f:
model_id = llm.model_id
logger.debug('Generating service vars file for %s at %s (dir=%s)', model_name, '_service_vars.py', llm_fs.getsyspath('/'))
with open(_service_vars_file.__fspath__(), 'r') as f:
src_contents = f.readlines()
for it in src_contents:
if OPENLLM_MODEL_NAME in it:
src_contents[src_contents.index(it)] = (ModelNameFormatter(model_name).vformat(it)[:-(len(OPENLLM_MODEL_NAME) + 3)] + '\n')
if OPENLLM_MODEL_ID in it:
src_contents[src_contents.index(it)] = (ModelIdFormatter(model_id).vformat(it)[:-(len(OPENLLM_MODEL_ID) + 3)] + '\n')
elif OPENLLM_MODEL_ADAPTER_MAP in it:
src_contents[src_contents.index(it)] = (ModelAdapterMapFormatter(orjson.dumps(adapter_map).decode()).vformat(it)[:-(len(OPENLLM_MODEL_ADAPTER_MAP) + 3)] + '\n')
script = f"# GENERATED BY 'openllm build {model_name}'. DO NOT EDIT\n\n" + ''.join(src_contents)
if DEBUG: logger.info('Generated script:\n%s', script)
llm_fs.writetext(llm.config['service_name'], script)
llm_fs.writetext('_service_vars.py', script)
logger.debug('Generating service file for %s at %s (dir=%s)', model_name, llm.config['service_name'], llm_fs.getsyspath('/'))
with open(_service_file.__fspath__(), 'r') as f:
service_src = f.read()
llm_fs.writetext(llm.config['service_name'], service_src)
@inject
def create_bento(bento_tag: bentoml.Tag,