File size: 3,876 Bytes
3824ea0 |
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 |
import os
import re
import json
from openai import OpenAI
from tqdm import tqdm
from rdkit import Chem
from rdkit import RDLogger
RDLogger.DisableLog('rdApp.*')
def canonicalize_smiles(smiles):
mol = Chem.MolFromSmiles(smiles)
return Chem.MolToSmiles(mol, canonical=True)
class Agent():
def __init__(self, url="http://localhost:8000/v1"):
self.client = OpenAI(
api_key="0",
base_url=url
)
print ("[Client Init] done")
def generate(self, query):
messages = [
{
"role": "user",
"content": query
}
]
result = self.client.chat.completions.create(
messages=messages,
model="general")
return result.choices[0].message.content
class AgentMolLM(Agent):
def __init__(self, url="http://localhost:8000/v1"):
super(AgentMolLM, self).__init__(url=url)
def generate(self, query, temperature=0.01, top_p=0.01, n_resp=1, add_no_think=False):
query = query+" /no_think" if add_no_think else query
messages = [
{
"role": "user",
"content": query
}
]
result = self.client.chat.completions.create(
messages=messages,
model="mol_lm",
temperature=temperature,
top_p=top_p,
n=n_resp)
if n_resp>1:
results = list()
for i in range(n_resp):
results.append(result.choices[i].message.content)
return results
else:
return result.choices[0].message.content
class ExtractorSMILES():
def __init__(self):
self.pattern = r"<SMILES>(.*?)</SMILES>"
def extract(self, text):
res = list()
for smi in re.findall(self.pattern, text, re.DOTALL):
res.append(smi.replace(";", "."))
return res
class ExtractorFormula():
def __init__(self):
self.pattern = r"<MOLFORMULA>(.*?)</MOLFORMULA>"
def extract(self, text):
return re.findall(self.pattern, text, re.DOTALL)
class ExtractorIUPAC():
def __init__(self):
self.pattern = r"<IUPAC>(.*?)</IUPAC>"
def extract(self, text):
return re.findall(self.pattern, text, re.DOTALL)
class ExtractorValue():
def __init__(self):
self.pattern = r"[+-]?\d*\.\d+"
def extract(self, text):
return re.findall(self.pattern, text)
class ExtractorYN():
def __init__(self):
pass
def extract(self, text):
if "yes" in text.lower():
return ["yes"]
elif "no" in text.lower():
return ["no"]
else:
return ["N/A"]
class ExtractorText():
def __init__(self):
pass
def extract(self, text):
return [text]
class Extractor():
def __init__(self, task_type="smiles"):
if task_type == "smiles":
self.extractor = ExtractorSMILES()
elif task_type == "formula":
self.extractor = ExtractorFormula()
elif task_type == "iupac":
self.extractor = ExtractorIUPAC()
elif task_type == "value":
self.extractor = ExtractorValue()
elif task_type == "bool":
self.extractor = ExtractorYN()
else:
self.extractor = ExtractorText()
def extract(self, text):
text = text.split("</think>\n")[-1]
resp = self.extractor.extract(text=text)
res = resp[0].strip() if len(resp) else ""
return res
if __name__ == "__main__":
agent = AgentMolLM()
QUERY = "你是谁"
ADD_NO_THINK = True
# resp = agent.generate(QUERY, add_no_think=ADD_NO_THINK)
resp = agent.generate(QUERY, add_no_think=ADD_NO_THINK, temperature=0.6, top_p=0.95, n_resp=2)
print (resp) |