Add api for sessions and add max_tokens for tenant_llm (#3472)

### What problem does this PR solve?

Add api for sessions and add max_tokens for tenant_llm

### Type of change

- [x] New Feature (non-breaking change which adds functionality)

---------

Co-authored-by: liuhua <10215101452@stu.ecun.edu.cn>

api/apps/sdk/session.py

@@ -13,21 +13,24 @@
 #  See the License for the specific language governing permissions and
 #  limitations under the License.
 #
+import re
 import json
 from functools import partial
 from uuid import uuid4
-
+from api.db import LLMType
 from flask import request, Response
-
+from api.db.services.dialog_service import ask
 from agent.canvas import Canvas
 from api.db import StatusEnum
 from api.db.db_models import API4Conversation
 from api.db.services.api_service import API4ConversationService
 from api.db.services.canvas_service import UserCanvasService
 from api.db.services.dialog_service import DialogService, ConversationService, chat
+from api.db.services.knowledgebase_service import KnowledgebaseService
 from api.utils import get_uuid
 from api.utils.api_utils import get_error_data_result
 from api.utils.api_utils import get_result, token_required
+from api.db.services.llm_service import LLMBundle
 
 
 @manager.route('/chats/<chat_id>/sessions', methods=['POST'])
@@ -342,7 +345,7 @@ def agent_completion(tenant_id, agent_id):
                 yield "data:" + json.dumps({"code": 500, "message": str(e),
                                             "data": {"answer": "**ERROR**: " + str(e), "reference": []}},
                                            ensure_ascii=False) + "\n\n"
-            yield "data:" + json.dumps({"code": 0, "message": "", "data": True}, ensure_ascii=False) + "\n\n"
+            yield "data:" + json.dumps({"code": 0, "data": True}, ensure_ascii=False) + "\n\n"
 
         resp = Response(sse(), mimetype="text/event-stream")
         resp.headers.add_header("Cache-control", "no-cache")
@@ -366,7 +369,7 @@ def agent_completion(tenant_id, agent_id):
 
 @manager.route('/chats/<chat_id>/sessions', methods=['GET'])
 @token_required
-def list(chat_id,tenant_id):
+def list_session(chat_id,tenant_id):
     if not DialogService.query(tenant_id=tenant_id, id=chat_id, status=StatusEnum.VALID.value):
         return get_error_data_result(message=f"You don't own the assistant {chat_id}.")
     id = request.args.get("id")
@@ -441,4 +444,80 @@ def delete(tenant_id,chat_id):
         if not conv:
             return get_error_data_result(message="The chat doesn't own the session")
         ConversationService.delete_by_id(id)
-    return get_result()
+    return get_result()
+
+@manager.route('/sessions/ask', methods=['POST'])
+@token_required
+def ask_about(tenant_id):
+    req = request.json
+    if not req.get("question"):
+        return get_error_data_result("`question` is required.")
+    if not req.get("dataset_ids"):
+        return get_error_data_result("`dataset_ids` is required.")
+    if not isinstance(req.get("dataset_ids"),list):
+        return get_error_data_result("`dataset_ids` should be a list.")
+    req["kb_ids"]=req.pop("dataset_ids")
+    for kb_id in req["kb_ids"]:
+        if not KnowledgebaseService.accessible(kb_id,tenant_id):
+            return get_error_data_result(f"You don't own the dataset {kb_id}.")
+        kbs = KnowledgebaseService.query(id=kb_id)
+        kb = kbs[0]
+        if kb.chunk_num == 0:
+            return get_error_data_result(f"The dataset {kb_id} doesn't own parsed file")
+    uid = tenant_id
+    def stream():
+        nonlocal req, uid
+        try:
+            for ans in ask(req["question"], req["kb_ids"], uid):
+                yield "data:" + json.dumps({"code": 0, "message": "", "data": ans}, ensure_ascii=False) + "\n\n"
+        except Exception as e:
+            yield "data:" + json.dumps({"code": 500, "message": str(e),
+                                        "data": {"answer": "**ERROR**: " + str(e), "reference": []}},
+                                       ensure_ascii=False) + "\n\n"
+        yield "data:" + json.dumps({"code": 0, "message": "", "data": True}, ensure_ascii=False) + "\n\n"
+
+    resp = Response(stream(), mimetype="text/event-stream")
+    resp.headers.add_header("Cache-control", "no-cache")
+    resp.headers.add_header("Connection", "keep-alive")
+    resp.headers.add_header("X-Accel-Buffering", "no")
+    resp.headers.add_header("Content-Type", "text/event-stream; charset=utf-8")
+    return resp
+
+
+@manager.route('/sessions/related_questions', methods=['POST'])
+@token_required
+def related_questions(tenant_id):
+    req = request.json
+    if not req.get("question"):
+        return get_error_data_result("`question` is required.")
+    question = req["question"]
+    chat_mdl = LLMBundle(tenant_id, LLMType.CHAT)
+    prompt = """
+Objective: To generate search terms related to the user's search keywords, helping users find more valuable information.
+Instructions:
+ - Based on the keywords provided by the user, generate 5-10 related search terms.
+ - Each search term should be directly or indirectly related to the keyword, guiding the user to find more valuable information.
+ - Use common, general terms as much as possible, avoiding obscure words or technical jargon.
+ - Keep the term length between 2-4 words, concise and clear.
+ - DO NOT translate, use the language of the original keywords.
+
+### Example:
+Keywords: Chinese football
+Related search terms:
+1. Current status of Chinese football
+2. Reform of Chinese football
+3. Youth training of Chinese football
+4. Chinese football in the Asian Cup
+5. Chinese football in the World Cup
+
+Reason:
+ - When searching, users often only use one or two keywords, making it difficult to fully express their information needs.
+ - Generating related search terms can help users dig deeper into relevant information and improve search efficiency. 
+ - At the same time, related terms can also help search engines better understand user needs and return more accurate search results.
+
+"""
+    ans = chat_mdl.chat(prompt, [{"role": "user", "content": f"""
+Keywords: {question}
+Related search terms:
+    """}], {"temperature": 0.9})
+    return get_result(data=[re.sub(r"^[0-9]\. ", "", a) for a in ans.split("\n") if re.match(r"^[0-9]\. ", a)])