Spaces:
Paused
Paused
File size: 10,118 Bytes
083f48d 89b86b6 083f48d 1511b30 f7e1b96 287f046 6e54404 287f046 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 |
import os
import openai
import gradio as gr
openai.api_key = os.environ["OpenAI_Key"]
HUGGINGFACE_TOKEN = os.environ["HF_KEY"]
# huggingface-cli login --token $HUGGINGFACE_TOKEN
from transformers import TFGPT2LMHeadModel, AutoTokenizer
tokenizer = AutoTokenizer.from_pretrained('Gracoy/ingredients_compatibility_GPT2_S')
model = TFGPT2LMHeadModel.from_pretrained('Gracoy/ingredients_compatibility_GPT2_S')
def assistant(assist_prompt):
response = openai.ChatCompletion.create(
model='gpt-3.5-turbo',
messages=assist_prompt,
temperature=0.7)
return response["choices"][0]["message"]["content"].strip()
def proper_ingredient(ingredients, meal):
if not ingredients:
return
assist_prompt1 = [{"role": "user", "content": f"列出三道使用{ingredients}裡食材適合當{meal}的料理，列舉即可不須詳細說明"}]
prompt = [
{"role": "system", "content": "你是一個烹飪經驗豐富的廚師，知道哪些食材適合拿來搭配來製作平易近人且美味的料理"},
{"role": "assistant", "content": assistant(assist_prompt1)},
{"role": "user", "content": f"請給我三組在{ingredients}中適合一起烹飪的食材組合，不需要料理名稱只要寫出食材和為何他們適合搭配在一起就好，\
列出的食材組合需至少包含一樣原本提供的食材，###輸出格式：至少一樣原本食材+新增食材：為何適合搭配在一起###"}
]
response = openai.ChatCompletion.create(
model='gpt-3.5-turbo',
messages=prompt,
temperature=0.7)
return response["choices"][0]["message"]["content"].strip()
def get_ingre(ingredients):
final_ingre = ingredients
return final_ingre
def gen_recipe(final_ingredients, flavor, meal):
global step
global round
global history
step = 9
round = 0
history = []
if not final_ingredients:
return
assist_prompt1 = [{"role": "user", "content": f"請用三句話說明料理中的{flavor}風味，列出{flavor}風味常使用的香料及調味料，列舉即可不須詳細說明"}]
assist_prompt2 = [{"role": "user", "content": f"列出三樣以{final_ingredients}裡的食材烹調出適合當{meal}的{flavor}風味料理，列舉即可不須詳細說明"}]
prompt = [
{"role": "system", "content": "你是一個烹飪經驗豐富的廚師，擅長使用手邊的食材和簡潔的烹飪步驟來製作平易近人的家庭料理"},
{"role": "assistant", "content": assistant(assist_prompt1)},
{"role": "assistant", "content": assistant(assist_prompt2)},
{"role": "user", "content": f"請在上述料理中選擇一個，並針對此料理給我一份步驟數在{step}步以內的食譜，烹飪方式不要用炒的，食材用量要詳細，\
食譜要包含用到的調味料，烹飪步驟要簡潔；###使用食材：{final_ingredients}，料理風味：{flavor}，餐點：{meal}\
###；###輸出格式:料理名稱,使用食材,使用調味料,烹飪步驟,分成三個區塊顯示###"}
]
response = openai.ChatCompletion.create(
model='gpt-3.5-turbo',
messages=prompt,
temperature=0.7)
round += 1
history.append(response["choices"][0]["message"]["content"].strip())
return response["choices"][0]["message"]["content"].strip()
def another_recipe(final_ingredients, flavor, meal):
global step
global round
global history
if round == 0 or (not final_ingredients):
return
step = 9
assist_prompt1 = [{"role": "user", "content": f"請用三句話說明料理中的{flavor}風味，列出{flavor}風味常使用的香料及調味料，列舉即可不須詳細說明"}]
assist_prompt2 = [{"role": "user", "content": f"你給我的{history[round-1]}料理名稱與食譜我都不喜歡，\
請列出其他三樣以{final_ingredients}裡的食材烹調出適合當{meal}的{flavor}風味料理，列舉即可不須詳細說明"}]
prompt = [
{"role": "system", "content": "你是一個烹飪經驗豐富的家庭主婦，擅長使用簡潔的烹飪步驟來製作平易近人的料理"},
{"role": "assistant", "content": assistant(assist_prompt1)},
{"role": "assistant", "content": assistant(assist_prompt2)},
{"role": "user", "content": f"請在上述料理中給我一份料理名稱不在{history}裡且步驟數在{step}步以內的食譜，烹飪方式不要用炒的, \
食材用量要詳細，食譜要包含用到的調味料，烹飪步驟的描述要簡潔；###使用食材：{final_ingredients}，料理風味：{flavor}，餐點：{meal}\
###；###輸出格式:料理名稱,使用食材,使用調味料,烹飪步驟,分成三個區塊顯示###"}
]
response = openai.ChatCompletion.create(
model='gpt-3.5-turbo',
messages=prompt,
temperature=0.5)
round += 1
history.append(response["choices"][0]["message"]["content"].strip())
return response["choices"][0]["message"]["content"].strip()
def simpler_recipe(final_ingredients, flavor, meal):
global step
global round
global history
if round == 0 or (not final_ingredients):
return
step = step - 2 if step > 5 else step
assist_prompt1 = [{"role": "user", "content": f"請用三句話說明料理中的{flavor}風味，列出{flavor}風味常使用的香料及調味料，列舉即可不須詳細說明"}]
assist_prompt2 = [{"role": "user", "content": f"你給我的{history[round-1]}食譜做法太難了，請列出其他三樣以{final_ingredients}裡的食材，\
烹調出適合當{meal}的{flavor}風味料理，列舉即可不須詳細說明"}]
prompt = [
{"role": "system", "content": "你是一個烹飪技巧不熟練的新手廚師，只能製作出簡單常見的料理"},
{"role": "assistant", "content": assistant(assist_prompt1)},
{"role": "assistant", "content": assistant(assist_prompt2)},
{"role": "user", "content": f"請給我一份料理名稱不在{history}裡且烹飪步驟在{step}步以內的食譜，先用一句話說明你對{history[round-1]}食譜做了哪些調整，烹飪方式不要用炒的，\
食材用量要詳細，最多只能使用三種調味料，烹飪步驟要簡潔；###使用食材：{final_ingredients}，料理風味：{flavor}，餐點：{meal}\
###；###輸出格式:料理名稱,使用食材,使用調味料,烹飪步驟,分成三個區塊顯示###"}
]
response = openai.ChatCompletion.create(
model='gpt-3.5-turbo',
messages=prompt,
temperature=0.7)
round += 1
history.append(response["choices"][0]["message"]["content"].strip())
return response["choices"][0]["message"]["content"].strip()
def compatibility_predict(ingredients):
response = openai.ChatCompletion.create(
model='gpt-3.5-turbo',
messages=[{"role": "user", "content": f"請把以下輸入翻譯成英文:{ingredients}"}]
)
translated = response['choices'][0]['message']['content'].strip().lower().replace('.', "")
prefix = 'I want to cook, please select the suitable ingredients from the following list: '
prompt = prefix + translated+ '.'
input = tokenizer('<BOS>' + prompt, return_tensors='tf')
encoded_outputs = model.generate(input['input_ids'], max_length=128, num_beams=3, num_return_sequences=1, early_stopping=True)
mid = 'The best combinations are: '
suffix = '. Enjoy your dish!'
res = []
for output in encoded_outputs:
ans = tokenizer.decode(output, skip_special_tokens=True)
ans = ans.replace(prompt, "").strip()
ans = ans.replace(mid, "").strip()
ans = ans.replace(suffix, "").strip()
ans = set(ans.split(', '))
res.append(list(ans))
response = openai.ChatCompletion.create(
model='gpt-3.5-turbo',
messages=[{"role": "user", "content": f"請把以下輸入翻譯成繁體中文，翻譯時請以台灣(Taiwan)的習慣用詞為準:{', '.join(ans)}"}]
)
return response["choices"][0]["message"]["content"].strip()
with gr.Blocks(css="footer{display: none !important;}", theme=gr.themes.Soft(primary_hue="zinc", neutral_hue="sky")) as demo1:
gr.Markdown(""" # 今晚我想來點 """)
with gr.Row():
with gr.Column():
ingredients = gr.Textbox(label='手邊有哪些其他食材呢？')
#meal = gr.Textbox(label='現在要吃的是早餐，午餐，點心，晚餐，還是宵夜？')
meal = gr.Radio(["早餐", "午餐", "晚餐", "點心", "宵夜"], label="你現在要烹煮的是哪一餐？")
submit_bnt = gr.Button('請推薦食材組合給我')
proper_set = gr.Textbox(label='推薦的食材組合')
with gr.Column():
final_ingredients = gr.Textbox(label="最後想使用的食材有哪些？", info="預設帶入原本輸入的食材，可再自行增減")
flavor = gr.Dropdown(['中式','日式','美式','韓式','義式','泰式','都可以'],label='你希望成品帶有哪些風味？', info="請從選單中任選一個")
submit_bnt1 = gr.Button('產生食譜吧！')
recipe = gr.Textbox(label='食譜')
with gr.Row():
simpler_btn = gr.Button('太難了')
another_btn = gr.Button('不喜歡')
submit_bnt.click(fn=compatibility_predict, inputs=[ingredients], outputs=[proper_set]) # 這邊把原本的proper_ingredient換成我們model的API, 其他按鈕都沒改
submit_bnt.click(fn=get_ingre, inputs=[ingredients], outputs=[final_ingredients])
submit_bnt1.click(fn=gen_recipe, inputs=[final_ingredients, flavor, meal], outputs=[recipe])
simpler_btn.click(fn=simpler_recipe, inputs=[final_ingredients, flavor, meal], outputs=[recipe])
another_btn.click(fn=another_recipe, inputs=[final_ingredients, flavor, meal], outputs=[recipe])
demo1.launch()
|