SoybeanMilk
commited on
Commit
•
aa0d25a
1
Parent(s):
5f1dbf2
Update app.py
Browse files
app.py
CHANGED
@@ -57,10 +57,10 @@ def cp_text(input_text):
|
|
57 |
def cp_clear():
|
58 |
pyperclip.clear()
|
59 |
|
60 |
-
# Split the text into
|
61 |
def process_text_input_text(input_text):
|
62 |
-
# Split the text into
|
63 |
-
chunks = [input_text[i:i+
|
64 |
return chunks
|
65 |
|
66 |
def process_and_translate(api_key, input_text, src_lang, tgt_lang):
|
@@ -73,22 +73,46 @@ def process_and_translate(api_key, input_text, src_lang, tgt_lang):
|
|
73 |
if chunk is None or chunk == "":
|
74 |
translated_chunks.append("System prompt: There is no content to translate!")
|
75 |
else:
|
76 |
-
prompt = f"This is an {src_lang} to {tgt_lang} translation, please provide the {tgt_lang} translation for this
|
|
|
|
|
77 |
genai.configure(api_key=api_key)
|
78 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
79 |
response = model.generate_content([prompt, chunk],
|
80 |
-
generation_config=
|
81 |
-
|
82 |
-
candidate_count=1,
|
83 |
-
max_output_tokens=2048,
|
84 |
-
temperature=0.3,
|
85 |
-
top_p=1,
|
86 |
-
)
|
87 |
)
|
88 |
translated_chunks.append(response.text)
|
89 |
|
90 |
# Join the translated chunks back together into a single string
|
91 |
-
response = ''.join(translated_chunks)
|
92 |
|
93 |
return response
|
94 |
|
@@ -104,20 +128,42 @@ def process_and_summary(api_key, input_text, src_lang, tgt_lang):
|
|
104 |
else:
|
105 |
prompt = f"This is an {src_lang} to {tgt_lang} summarization and knowledge key points, please provide the {tgt_lang} summarization and list the {tgt_lang} knowledge key points for this sentence. Do not provide any explanations or text apart from the summarization.\n{src_lang}: "
|
106 |
genai.configure(api_key=api_key)
|
107 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
108 |
response = model.generate_content([prompt, chunk],
|
109 |
-
generation_config=
|
110 |
-
|
111 |
-
candidate_count=1,
|
112 |
-
max_output_tokens=2048,
|
113 |
-
temperature=0.3,
|
114 |
-
top_p=1,
|
115 |
-
)
|
116 |
)
|
117 |
translated_chunks.append(response.text)
|
118 |
|
119 |
# Join the translated chunks back together into a single string
|
120 |
-
response = '\n
|
121 |
|
122 |
return response
|
123 |
|
@@ -199,6 +245,7 @@ def main():
|
|
199 |
# ---------------------- 复制到剪贴板 ----------------------
|
200 |
cp_btn.click(fn=cp_text, inputs=[outputs_tr_text], outputs=[])
|
201 |
cp_clear_btn.click(fn=cp_clear, inputs=[], outputs=[])
|
|
|
202 |
|
203 |
ocr_tr.launch(inbrowser=True)
|
204 |
|
|
|
57 |
def cp_clear():
|
58 |
pyperclip.clear()
|
59 |
|
60 |
+
# Split the text into 2000 character chunks
|
61 |
def process_text_input_text(input_text):
|
62 |
+
# Split the text into 2000 character chunks
|
63 |
+
chunks = [input_text[i:i+2000] for i in range(0, len(input_text), 2000)]
|
64 |
return chunks
|
65 |
|
66 |
def process_and_translate(api_key, input_text, src_lang, tgt_lang):
|
|
|
73 |
if chunk is None or chunk == "":
|
74 |
translated_chunks.append("System prompt: There is no content to translate!")
|
75 |
else:
|
76 |
+
prompt = f"This is an {src_lang} to {tgt_lang} translation, please provide the {tgt_lang} translation for this paragraph. Do not provide any explanations or text apart from the translation.\n{src_lang}: "
|
77 |
+
#prompt = f"This is an {src_lang} to {tgt_lang} translation, please provide the {tgt_lang} translation for this sentence. Do not provide any explanations or text apart from the translation.\n{src_lang}: "
|
78 |
+
|
79 |
genai.configure(api_key=api_key)
|
80 |
+
|
81 |
+
generation_config = {
|
82 |
+
"candidateCount": 1,
|
83 |
+
"maxOutputTokens": 2048,
|
84 |
+
"temperature": 0.3,
|
85 |
+
"topP": 1
|
86 |
+
}
|
87 |
+
|
88 |
+
safety_settings = [
|
89 |
+
{
|
90 |
+
"category": "HARM_CATEGORY_HARASSMENT",
|
91 |
+
"threshold": "BLOCK_NONE",
|
92 |
+
},
|
93 |
+
{
|
94 |
+
"category": "HARM_CATEGORY_HATE_SPEECH",
|
95 |
+
"threshold": "BLOCK_NONE",
|
96 |
+
},
|
97 |
+
{
|
98 |
+
"category": "HARM_CATEGORY_SEXUALLY_EXPLICIT",
|
99 |
+
"threshold": "BLOCK_NONE",
|
100 |
+
},
|
101 |
+
{
|
102 |
+
"category": "HARM_CATEGORY_DANGEROUS_CONTENT",
|
103 |
+
"threshold": "BLOCK_NONE",
|
104 |
+
},
|
105 |
+
]
|
106 |
+
|
107 |
+
model = genai.GenerativeModel(model_name='gemini-pro')
|
108 |
response = model.generate_content([prompt, chunk],
|
109 |
+
#generation_config=generation_config,
|
110 |
+
safety_settings=safety_settings
|
|
|
|
|
|
|
|
|
|
|
111 |
)
|
112 |
translated_chunks.append(response.text)
|
113 |
|
114 |
# Join the translated chunks back together into a single string
|
115 |
+
response = '\n\n'.join(translated_chunks)
|
116 |
|
117 |
return response
|
118 |
|
|
|
128 |
else:
|
129 |
prompt = f"This is an {src_lang} to {tgt_lang} summarization and knowledge key points, please provide the {tgt_lang} summarization and list the {tgt_lang} knowledge key points for this sentence. Do not provide any explanations or text apart from the summarization.\n{src_lang}: "
|
130 |
genai.configure(api_key=api_key)
|
131 |
+
|
132 |
+
generation_config = {
|
133 |
+
"candidateCount": 1,
|
134 |
+
"maxOutputTokens": 2048,
|
135 |
+
"temperature": 0.3,
|
136 |
+
"topP": 1
|
137 |
+
}
|
138 |
+
|
139 |
+
safety_settings = [
|
140 |
+
{
|
141 |
+
"category": "HARM_CATEGORY_HARASSMENT",
|
142 |
+
"threshold": "BLOCK_NONE",
|
143 |
+
},
|
144 |
+
{
|
145 |
+
"category": "HARM_CATEGORY_HATE_SPEECH",
|
146 |
+
"threshold": "BLOCK_NONE",
|
147 |
+
},
|
148 |
+
{
|
149 |
+
"category": "HARM_CATEGORY_SEXUALLY_EXPLICIT",
|
150 |
+
"threshold": "BLOCK_NONE",
|
151 |
+
},
|
152 |
+
{
|
153 |
+
"category": "HARM_CATEGORY_DANGEROUS_CONTENT",
|
154 |
+
"threshold": "BLOCK_NONE",
|
155 |
+
},
|
156 |
+
]
|
157 |
+
|
158 |
+
model = genai.GenerativeModel(model_name='gemini-pro')
|
159 |
response = model.generate_content([prompt, chunk],
|
160 |
+
#generation_config=generation_config,
|
161 |
+
safety_settings=safety_settings
|
|
|
|
|
|
|
|
|
|
|
162 |
)
|
163 |
translated_chunks.append(response.text)
|
164 |
|
165 |
# Join the translated chunks back together into a single string
|
166 |
+
response = '\n\n*Next Paragraph*\n\n'.join(translated_chunks)
|
167 |
|
168 |
return response
|
169 |
|
|
|
245 |
# ---------------------- 复制到剪贴板 ----------------------
|
246 |
cp_btn.click(fn=cp_text, inputs=[outputs_tr_text], outputs=[])
|
247 |
cp_clear_btn.click(fn=cp_clear, inputs=[], outputs=[])
|
248 |
+
|
249 |
|
250 |
ocr_tr.launch(inbrowser=True)
|
251 |
|