update app.py

app.py

@@ -121,7 +121,7 @@ examples = [
     ],
     [
         "receipt.png",
-        "เบอร์โทรร้านอะไรคะ"
+        "เบอร์โทรร้านอะไรคะ?"
     ],
     [
         "poster.png",
@@ -166,7 +166,7 @@ prompt_files = {
     "Which quality component has the icon of a pen in it?": "website_design_guide.jpeg",
     "Which gift idea needs a printer?": "poster.png",
     "患者さんは何でお金を払いますか。": "medical_bill_2.jpg",
-    "เบอร์โทรร้านอะไรคะ": "receipt.png",
+    "เบอร์โทรร้านอะไรคะ?": "receipt.png",
     "试卷当前部分考察什么内容？": "exam.png",
     "For Rage, who is the author listed as? ": "book.png",
 }
@@ -237,11 +237,22 @@ def np2base64(image_np):
     return base64_str
 
 
-def img2base64(path):
+def get_base64(path):
+    if path.startswith("http://") or path.startswith("https://"):
+        resp = requests.get(path, allow_redirects=True, stream=True)
+        b = resp.raw
+    else:
+        b = open(path, "rb")
+
     if path.endswith(".pdf"):
-        base64_str = np2base64(load_document(path)[0])
+        images_list = []
+        pdfreader = PdfReader(stream=b.read())
+        for p_no in range(0, min(pdfreader._inpdf.page_count, 1)):
+            img_np = pdfreader.get_page_image(pageno=p_no)
+            images_list.append(img_np)
+        base64_str = np2base64(images_list[0])
     else:
-        base64_str = base64.b64encode(open(path, 'rb').read()).decode()
+        base64_str = base64.b64encode(b.read()).decode()
     return base64_str
 
 
@@ -255,7 +266,7 @@ def process_prompt(prompt, document, lang="ch"):
 
     url = f"https://aip.baidubce.com/rpc/2.0/nlp-itec/poc/docprompt?access_token={access_token}"
     
-    base64_str = img2base64(document)
+    base64_str = get_base64(document)
 
     r = requests.post(url, json={"doc": base64_str, "prompt": [prompt], "lang": lang})
     response = r.json()