用 AI 實際解決工作問題！這種讓人痛苦的工作，還是讓AI來吧

任務說明

背景 (涉及一點技術上的名詞)

來點 AI ！

'''
1 接續前面步驟，讀取word檔並偵測出表格
2 回傳結果tables
'''
print(tables[3]) # 查看眾多tables中的第4個table，即本文的表1

[['Item', 'Unit ', 'Limit of Range'],  
 ['Total Moisture (TM) \n(As received basis)', '% WT', 'Max. 28'],  
 ['Volatile Matter (Air dried basis)\nFixed Carbon (Air dried basis)\nAsh (Air dried basis)',   '% WT\n% WT\n% WT',   'Min. 22∼Max. 45\nMax. 60\nMax. 17'],  
 ['Total Sulfur (As received basis)', '% WT', 'Max. 0.6'],  
 ['Grindability (HGI)', '', 'Min. 45~Max. 65'],  
 ['(As Received Basis)\nGross Calorific Value\nNet Calorific Value *',   '\nkcal/kg\nkcal/kg',   '\n\nMin. 4,750'],  
 ['Nitrogen (Dry ash free basis)', '% WT', 'Max. 2.2'],  
 ['Ash Fusion Temperature (IDT) \n(Reducing Atmosphere)  ',   'oC',   'Min. 1,150'],  
 ['CSN(Crucible Swelling Number)', '', 'Less than 2\n(0≤CSN<2)'],  
 ['Mercury(HG)', '㎍/g', ''],  
 ['Ash Analysis\nSiO₂\nFe₂O₃\nNa₂O\nK₂O',   '\n% WT\n% WT\n% WT\n% WT',   '\nMax. 70\nMax. 16\nMax. 2\nMax. 3'],  
 ['Size Distribution\nAbove  50mm\nUnder  2mm \nUnder  0.25mm',   '\n%\n%\n%',   '\nMax. 5\nMax. 40\nMax. 17']
]

import openai
openai.api_key = TOKEN # 輸入你的API KEY
tb_specs = tables[3] # 本文的表1
output_txt = ""
for sub_lst in tb_specs:
　　for ele in sub_lst:
　　　　output_txt += " " + ele.strip("\n")
print(output_txt)

' Item Unit  Limit of Range Total Moisture (TM) \n(As received basis) % WT Max. 28 Volatile Matter (Air dried basis)\nFixed Carbon (Air dried basis)\nAsh (Air dried basis) % WT\n% WT\n% WT Min. 22∼Max. 45\nMax. 60\nMax. 17 Total Sulfur (As received basis) % WT Max. 0.6 Grindability (HGI)  Min. 45~Max. 65 (As Received Basis)\nGross Calorific Value\nNet Calorific Value * kcal/kg\nkcal/kg Min. 4,750 Nitrogen (Dry ash free basis) % WT Max. 2.2 Ash Fusion Temperature (IDT) \n(Reducing Atmosphere)   oC Min. 1,150 CSN(Crucible Swelling Number)  Less than 2\n(0≤CSN<2) Mercury(HG) ㎍/g  Ash Analysis\nSiO₂\nFe₂O₃\nNa₂O\nK₂O % WT\n% WT\n% WT\n% WT Max. 70\nMax. 16\nMax. 2\nMax. 3 Size Distribution\nAbove  50mm\nUnder  2mm \nUnder  0.25mm %\n%\n% Max. 5\nMax. 40\nMax. 17'

input = "請將內容整理成表格, 並告訴我 Net Calorific Value的數值, 回傳數字(要包含Min, Max)但不要回傳出表格"
completion = openai.ChatCompletion.create(
　　model="gpt-4",
　　messages=[
　　　　{"role": "system", "content": input},
　　　　{"role": "user", "content": output_txt},
　　]
)

print(completion.choices[0].message.content)
'Net Calorific Value 的數值範圍是：最小值 4,750 kcal/kg，最大值未提供。'

input = "請將內容整理成表格, 並告訴我 Total Sulfur (As received basis) 的數值, 回傳數字(要包含Min, Max, 如果有的話)但不要回傳出表格"
completion = openai.ChatCompletion.create(
model="gpt-4",
messages=[
　　{"role": "system", "content": input},
　　{"role": "user", "content": output_txt},
　]
)
print(completion.choices[0].message.content)
'Total Sulfur (As received basis) 的數值為 % WT Max. 0.6。'

兩種後續嘗試

三個心得