Feat: add vision LLM PDF parser (#6173)

### What problem does this PR solve?

Add vision LLM PDF parser

### Type of change

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

---------

Co-authored-by: Kevin Hu <kevinhu.sh@gmail.com>

rag/app/picture.py

@@ -21,8 +21,9 @@ from PIL import Image
 
 from api.db import LLMType
 from api.db.services.llm_service import LLMBundle
-from rag.nlp import tokenize
 from deepdoc.vision import OCR
+from rag.nlp import tokenize
+from rag.utils import clean_markdown_block
 
 ocr = OCR()
 
@@ -57,3 +58,32 @@ def chunk(filename, binary, tenant_id, lang, callback=None, **kwargs):
         callback(prog=-1, msg=str(e))
 
     return []
+
+
+def vision_llm_chunk(binary, vision_model, prompt=None, callback=None):
+    """
+    A simple wrapper to process image to markdown texts via VLM.
+
+    Returns:
+        Simple markdown texts generated by VLM.
+    """
+    callback = callback or (lambda prog, msg: None)
+
+    img = binary
+    txt = ""
+
+    try:
+        img_binary = io.BytesIO()
+        img.save(img_binary, format='JPEG')
+        img_binary.seek(0)
+
+        ans = clean_markdown_block(vision_model.describe_with_prompt(img_binary.read(), prompt))
+
+        txt += "\n" + ans
+
+        return txt
+
+    except Exception as e:
+        callback(-1, str(e))
+
+    return []