Don't release full image (#10654)

### What problem does this PR solve?

Introduced gpu profile in .env
Added Dockerfile_tei
fix datrie
Removed LIGHTEN flag

### Type of change

- [x] Documentation Update
- [x] Refactoring

rag/svr/jina_server.py

@@ -1,109 +0,0 @@
-#
-#  Copyright 2025 The InfiniFlow Authors. All Rights Reserved.
-#
-#  Licensed under the Apache License, Version 2.0 (the "License");
-#  you may not use this file except in compliance with the License.
-#  You may obtain a copy of the License at
-#
-#      http://www.apache.org/licenses/LICENSE-2.0
-#
-#  Unless required by applicable law or agreed to in writing, software
-#  distributed under the License is distributed on an "AS IS" BASIS,
-#  WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-#  See the License for the specific language governing permissions and
-#  limitations under the License.
-#
-
-from jina import Deployment
-from docarray import BaseDoc
-from jina import Executor, requests
-from transformers import AutoModelForCausalLM, AutoTokenizer, GenerationConfig
-import argparse
-import torch
-
-
-class Prompt(BaseDoc):
-    message: list[dict]
-    gen_conf: dict
-
-
-class Generation(BaseDoc):
-    text: str
-
-
-tokenizer = None
-model_name = ""
-
-
-class TokenStreamingExecutor(Executor):
-    def __init__(self, **kwargs):
-        super().__init__(**kwargs)
-        self.model = AutoModelForCausalLM.from_pretrained(
-            model_name, device_map="auto", torch_dtype="auto"
-        )
-
-    @requests(on="/chat")
-    async def generate(self, doc: Prompt, **kwargs) -> Generation:
-        text = tokenizer.apply_chat_template(
-            doc.message,
-            tokenize=False,
-        )
-        inputs = tokenizer([text], return_tensors="pt")
-        generation_config = GenerationConfig(
-            **doc.gen_conf,
-            eos_token_id=tokenizer.eos_token_id,
-            pad_token_id=tokenizer.eos_token_id
-        )
-        generated_ids = self.model.generate(
-            inputs.input_ids, generation_config=generation_config
-        )
-        generated_ids = [
-            output_ids[len(input_ids) :]
-            for input_ids, output_ids in zip(inputs.input_ids, generated_ids)
-        ]
-
-        response = tokenizer.batch_decode(generated_ids, skip_special_tokens=True)[0]
-        yield Generation(text=response)
-
-    @requests(on="/stream")
-    async def task(self, doc: Prompt, **kwargs) -> Generation:
-        text = tokenizer.apply_chat_template(
-            doc.message,
-            tokenize=False,
-        )
-        input = tokenizer([text], return_tensors="pt")
-        input_len = input["input_ids"].shape[1]
-        max_new_tokens = 512
-        if "max_new_tokens" in doc.gen_conf:
-            max_new_tokens = doc.gen_conf.pop("max_new_tokens")
-        generation_config = GenerationConfig(
-            **doc.gen_conf,
-            eos_token_id=tokenizer.eos_token_id,
-            pad_token_id=tokenizer.eos_token_id
-        )
-        for _ in range(max_new_tokens):
-            output = self.model.generate(
-                **input, max_new_tokens=1, generation_config=generation_config
-            )
-            if output[0][-1] == tokenizer.eos_token_id:
-                break
-            yield Generation(
-                text=tokenizer.decode(output[0][input_len:], skip_special_tokens=True)
-            )
-            input = {
-                "input_ids": output,
-                "attention_mask": torch.ones(1, len(output[0])),
-            }
-
-
-if __name__ == "__main__":
-    parser = argparse.ArgumentParser()
-    parser.add_argument("--model_name", type=str, help="Model name or path")
-    parser.add_argument("--port", default=12345, type=int, help="Jina serving port")
-    args = parser.parse_args()
-    model_name = args.model_name
-    tokenizer = AutoTokenizer.from_pretrained(args.model_name)
-    with Deployment(
-        uses=TokenStreamingExecutor, port=args.port, protocol="grpc"
-    ) as dep:
-        dep.block()

rag/svr/task_executor.py

@@ -29,6 +29,7 @@ from api.db.services.pipeline_operation_log_service import PipelineOperationLogS
 from api.utils.api_utils import timeout
 from api.utils.base64_image import image2id
 from api.utils.log_utils import init_root_logger, get_project_base_directory
+from api.utils.configs import show_configs
 from graphrag.general.index import run_graphrag_for_kb
 from graphrag.utils import get_llm_cache, set_llm_cache, get_tags_from_cache, set_tags_to_cache
 from rag.flow.pipeline import Pipeline
@@ -475,7 +476,7 @@ async def embedding(docs, mdl, parser_config=None, callback=None):
     tk_count = 0
     if len(tts) == len(cnts):
         vts, c = await trio.to_thread.run_sync(lambda: mdl.encode(tts[0: 1]))
-        tts = np.concatenate([vts for _ in range(len(tts))], axis=0)
+        tts = np.concatenate([vts[0] for _ in range(len(tts))], axis=0)
         tk_count += c
 
     @timeout(60)
@@ -1061,7 +1062,10 @@ async def main():
           /____/
     """)
     logging.info(f'RAGFlow version: {get_ragflow_version()}')
+    show_configs()
     settings.init_settings()
+    from api.settings import EMBEDDING_CFG
+    logging.info(f'api.settings.EMBEDDING_CFG: {EMBEDDING_CFG}')
     print_rag_settings()
     if sys.platform != "win32":
         signal.signal(signal.SIGUSR1, start_tracemalloc_and_snapshot)