diff --git a/.gitignore b/.gitignore
index 93c4cb48c..d62218e87 100644
--- a/.gitignore
+++ b/.gitignore
@@ -6,7 +6,7 @@ __pycache__/
hudet/
cv/
layout_app.py
-resume/
+api/flask_session
# Remove Cargo.lock from gitignore if creating an executable, leave it for libraries
# More information here https://doc.rust-lang.org/cargo/guide/cargo-toml-vs-cargo-lock.html
diff --git a/api/apps/conversation_app.py b/api/apps/conversation_app.py
index 0d392bccf..f4d327993 100644
--- a/api/apps/conversation_app.py
+++ b/api/apps/conversation_app.py
@@ -163,6 +163,7 @@ def completion():
del req["conversation_id"]
del req["messages"]
ans = chat(dia, msg, **req)
+ if not conv.reference: conv.reference = []
conv.reference.append(ans["reference"])
conv.message.append({"role": "assistant", "content": ans["answer"]})
ConversationService.update_by_id(conv.id, conv.to_dict())
diff --git a/api/apps/dialog_app.py b/api/apps/dialog_app.py
index cc6f9810a..fccc0ecbb 100644
--- a/api/apps/dialog_app.py
+++ b/api/apps/dialog_app.py
@@ -32,7 +32,6 @@ def set_dialog():
dialog_id = req.get("dialog_id")
name = req.get("name", "New Dialog")
description = req.get("description", "A helpful Dialog")
- language = req.get("language", "Chinese")
top_n = req.get("top_n", 6)
similarity_threshold = req.get("similarity_threshold", 0.1)
vector_similarity_weight = req.get("vector_similarity_weight", 0.3)
@@ -80,7 +79,6 @@ def set_dialog():
"name": name,
"kb_ids": req["kb_ids"],
"description": description,
- "language": language,
"llm_id": llm_id,
"llm_setting": llm_setting,
"prompt_config": prompt_config,
diff --git a/api/apps/document_app.py b/api/apps/document_app.py
index 0649adaaa..9ce7e2b11 100644
--- a/api/apps/document_app.py
+++ b/api/apps/document_app.py
@@ -272,7 +272,9 @@ def get(doc_id):
response = flask.make_response(MINIO.get(doc.kb_id, doc.location))
ext = re.search(r"\.([^.]+)$", doc.name)
if ext:
- response.headers.set('Content-Type', 'application/%s'%ext.group(1))
+ if doc.type == FileType.VISUAL.value:
+ response.headers.set('Content-Type', 'image/%s'%ext.group(1))
+ else: response.headers.set('Content-Type', 'application/%s'%ext.group(1))
return response
except Exception as e:
return server_error_response(e)
diff --git a/api/db/db_models.py b/api/db/db_models.py
index 49aa169cf..9b2dc7ffb 100644
--- a/api/db/db_models.py
+++ b/api/db/db_models.py
@@ -464,6 +464,7 @@ class Knowledgebase(DataBaseModel):
avatar = TextField(null=True, help_text="avatar base64 string")
tenant_id = CharField(max_length=32, null=False)
name = CharField(max_length=128, null=False, help_text="KB name", index=True)
+ language = CharField(max_length=32, null=True, default="Chinese", help_text="English|Chinese")
description = TextField(null=True, help_text="KB description")
embd_id = CharField(max_length=128, null=False, help_text="default embedding model ID")
permission = CharField(max_length=16, null=False, help_text="me|team", default="me")
diff --git a/api/db/services/llm_service.py b/api/db/services/llm_service.py
index 6bc1150d7..2a9647402 100644
--- a/api/db/services/llm_service.py
+++ b/api/db/services/llm_service.py
@@ -57,7 +57,7 @@ class TenantLLMService(CommonService):
@classmethod
@DB.connection_context()
- def model_instance(cls, tenant_id, llm_type, llm_name=None):
+ def model_instance(cls, tenant_id, llm_type, llm_name=None, lang="Chinese"):
e, tenant = TenantService.get_by_id(tenant_id)
if not e:
raise LookupError("Tenant not found")
@@ -87,7 +87,7 @@ class TenantLLMService(CommonService):
if model_config["llm_factory"] not in CvModel:
return
return CvModel[model_config["llm_factory"]](
- model_config["api_key"], model_config["llm_name"])
+ model_config["api_key"], model_config["llm_name"], lang)
if llm_type == LLMType.CHAT.value:
if model_config["llm_factory"] not in ChatModel:
@@ -120,11 +120,11 @@ class TenantLLMService(CommonService):
class LLMBundle(object):
- def __init__(self, tenant_id, llm_type, llm_name=None):
+ def __init__(self, tenant_id, llm_type, llm_name=None, lang="Chinese"):
self.tenant_id = tenant_id
self.llm_type = llm_type
self.llm_name = llm_name
- self.mdl = TenantLLMService.model_instance(tenant_id, llm_type, llm_name)
+ self.mdl = TenantLLMService.model_instance(tenant_id, llm_type, llm_name, lang=lang)
assert self.mdl, "Can't find mole for {}/{}/{}".format(tenant_id, llm_type, llm_name)
def encode(self, texts: list, batch_size=32):
diff --git a/api/db/services/task_service.py b/api/db/services/task_service.py
index 87e84a12a..a63130150 100644
--- a/api/db/services/task_service.py
+++ b/api/db/services/task_service.py
@@ -27,7 +27,24 @@ class TaskService(CommonService):
@classmethod
@DB.connection_context()
def get_tasks(cls, tm, mod=0, comm=1, items_per_page=64):
- fields = [cls.model.id, cls.model.doc_id, cls.model.from_page,cls.model.to_page, Document.kb_id, Document.parser_id, Document.parser_config, Document.name, Document.type, Document.location, Document.size, Knowledgebase.tenant_id, Tenant.embd_id, Tenant.img2txt_id, Tenant.asr_id, cls.model.update_time]
+ fields = [
+ cls.model.id,
+ cls.model.doc_id,
+ cls.model.from_page,
+ cls.model.to_page,
+ Document.kb_id,
+ Document.parser_id,
+ Document.parser_config,
+ Document.name,
+ Document.type,
+ Document.location,
+ Document.size,
+ Knowledgebase.tenant_id,
+ Knowledgebase.language,
+ Tenant.embd_id,
+ Tenant.img2txt_id,
+ Tenant.asr_id,
+ cls.model.update_time]
docs = cls.model.select(*fields) \
.join(Document, on=(cls.model.doc_id == Document.id)) \
.join(Knowledgebase, on=(Document.kb_id == Knowledgebase.id)) \
@@ -42,7 +59,6 @@ class TaskService(CommonService):
.paginate(1, items_per_page)
return list(docs.dicts())
-
@classmethod
@DB.connection_context()
def do_cancel(cls, id):
@@ -54,12 +70,11 @@ class TaskService(CommonService):
pass
return True
-
@classmethod
@DB.connection_context()
def update_progress(cls, id, info):
- cls.model.update(progress_msg=cls.model.progress_msg + "\n"+info["progress_msg"]).where(
+ cls.model.update(progress_msg=cls.model.progress_msg + "\n" + info["progress_msg"]).where(
cls.model.id == id).execute()
if "progress" in info:
cls.model.update(progress=info["progress"]).where(
- cls.model.id == id).execute()
+ cls.model.id == id).execute()
diff --git a/api/flask_session/2029240f6d1128be89ddc32729463129 b/api/flask_session/2029240f6d1128be89ddc32729463129
deleted file mode 100644
index 60b84f8bf..000000000
Binary files a/api/flask_session/2029240f6d1128be89ddc32729463129 and /dev/null differ
diff --git a/api/utils/file_utils.py b/api/utils/file_utils.py
index b504cfef3..1d6f15f33 100644
--- a/api/utils/file_utils.py
+++ b/api/utils/file_utils.py
@@ -167,7 +167,11 @@ def thumbnail(filename, blob):
return "data:image/png;base64," + base64.b64encode(buffered.getvalue()).decode("utf-8")
if re.match(r".*\.(jpg|jpeg|png|tif|gif|icon|ico|webp)$", filename):
- return ("data:image/%s;base64,"%filename.split(".")[-1]) + base64.b64encode(Image.open(BytesIO(blob)).thumbnail((30, 30)).tobytes()).decode("utf-8")
+ image = Image.open(BytesIO(blob))
+ image.thumbnail((30, 30))
+ buffered = BytesIO()
+ image.save(buffered, format="png")
+ return "data:image/png;base64," + base64.b64encode(buffered.getvalue()).decode("utf-8")
if re.match(r".*\.(ppt|pptx)$", filename):
import aspose.slides as slides
diff --git a/deepdoc/README.md b/deepdoc/README.md
new file mode 100644
index 000000000..b2ad032dc
--- /dev/null
+++ b/deepdoc/README.md
@@ -0,0 +1,82 @@
+English | [简体中文](./README_zh.md)
+
+#*Deep*Doc
+
+---
+
+- [1. Introduction](#1)
+- [2. Vision](#2)
+- [3. Parser](#3)
+
+
+## 1. Introduction
+
+---
+With a bunch of documents from various domains with various formats and along with diverse retrieval requirements,
+an accurate analysis becomes a very challenge task. *Deep*Doc is born for that purpose.
+There 2 parts in *Deep*Doc so far: vision and parser.
+
+
+## 2. Vision
+
+---
+
+We use vision information to resolve problems as human being.
+ - OCR. Since a lot of documents presented as images or at least be able to transform to image,
+ OCR is a very essential and fundamental or even universal solution for text extraction.
+
+
+
+
+
+
+
+