File size: 11,143 Bytes
a12c07f |
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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 |
'''
This is the helper py file to save all kinds of ParsedAnswer type.
ParsedAnswer:
Input: language model raw output(words)
Output: formatted output according to the given pattern
Input_instrcution: There will be a few lines that could be imported to the model
as a in-context guidance
'''
import re
from .common import ParsedAnswer
from .exceptions import *
from typing import List
class LeftOrRight(ParsedAnswer):
def __init__(self, l_or_r:str, gpt_raw:str=None):
if l_or_r.lower().strip() not in ('left', 'right'):
# print("@@@@@@@@@@@@@@@@@@@@@@@@@@@@@")
# print(gpt_raw)
# print("@@@@@@@@@@@@@@@@@@@@@@@@@@@@@")
raise GPTOutputParseException("output of LLM should either be 'left' or 'right'")
self.data = l_or_r.lower().strip()
self.raw = gpt_raw
@staticmethod
def remove_answer_text(s:str):
# Define the regex pattern to match Python code blocks in Markdown
pattern = r'```\s*([^`]+)```'
# Use re.findall to extract all matches of the pattern
matches = re.findall(pattern, s, re.DOTALL)
# Join the matches into a single string
answer_text = '\n'.join(matches)
return answer_text
@staticmethod
def parser(gpt_raw:str) -> "LeftOrRight":
"""
@GPT4-doc-begin
gpt_raw: A block of text that has a single word in a text block, indicated by ```.
The word should be either "right" or "left".
For example,
The question that you've provided asks me to choose the image (left or right) that best aligns with the user text prompt.
Though the sample on the left is more realistic, the sample on the right is better aligned with the user prompt.
```
right
```
@GPT4-doc-end
"""
left_or_right = LeftOrRight.remove_answer_text(gpt_raw)
return LeftOrRight(left_or_right, gpt_raw=gpt_raw)
def __str__(self):
return self.data
class StarredList(ParsedAnswer):
def __init__(self, list_items:List[str], gpt_raw:str=None):
self.list_items = list_items
self.raw = gpt_raw
@staticmethod
def parse_bullet_points(input_string):
lines = input_string.split("\n")
bullet_points = []
current_bullet_point = ""
for line in lines:
if line.startswith("* "):
if current_bullet_point:
bullet_points.append(current_bullet_point.strip())
current_bullet_point = ""
current_bullet_point += line[2:] + " "
elif current_bullet_point:
current_bullet_point += line + " "
if current_bullet_point:
bullet_points.append(current_bullet_point.strip())
return bullet_points
@staticmethod
def parser(gpt_raw:str) -> "StarredList":
"""
@GPT4-doc-begin
gpt_raw: A new-line separated bulletpoint list that follows the following format:
Example:
* first item
* second item
...etc
@GPT4-doc-end
"""
list_items = StarredList.parse_bullet_points(gpt_raw)
return StarredList(list_items=list_items, gpt_raw=gpt_raw)
class PythonExecutableDiffAnswer(ParsedAnswer):
""" Code (python) difference
"""
def __init__(self, code_before, code_after, gpt_raw:str=None):
self.code_from = code_before
self.code_to = code_after
self.raw = gpt_raw
@staticmethod
def parser(gpt_raw:str) -> "PythonExecutableDiffAnswer":
"""
@GPT4-doc-begin
gpt_raw: A desired code change to a single line,
indicated by "Before:" and "After:" labels,
followed by python code blocks that indicate which line should be
changed, and to what.
Example:
Before:
```python
a = 1
```
After:
```python
a = 2
```
@GPT4-doc-end
"""
before_after = gpt_raw.split("After")
before_text = before_after[0]
assert "Before" in before_text
after_text = before_after[-1]
try:
before_string = PythonExecutableAnswer.remove_markdown_code(before_text)
after_string = PythonExecutableAnswer.remove_markdown_code(after_text)
except:
# print("@@@@@@@@@@@@@@@@@@@@@@@@@@@@@")
# print(gpt_raw)
# print("@@@@@@@@@@@@@@@@@@@@@@@@@@@@@")
raise GPTOutputParseException(f"Invalid input passsed into parsing function. The following could not be parsed:\n{gpt_raw}")
return PythonExecutableDiffAnswer(code_before=before_string, code_after=after_string, gpt_raw=gpt_raw)
def __str__(self):
return str(self.code)
class PythonExecutableAnswer(ParsedAnswer):
""" Code (Python)
"""
def __init__(self, code, gpt_raw:str=None):
self.code = code
self.raw = gpt_raw
@staticmethod
def remove_markdown_code(s:str):
if "```python" not in s:
raise GPTOutputParseException("No ```python found in the input.")
chunk1 = s.split("```python")[-1]
if "```" not in chunk1:
raise GPTOutputParseException("No ``` found to close the code block.")
python_code = chunk1.split("```")[0]
return python_code
@staticmethod
def parser(gpt_raw:str) -> "PythonExecutableAnswer":
"""
@GPT4-doc-begin
gpt_raw: string that contains a block of code to be executed in python.
you should add comments to explain what you are doing.
For instance,
We will create a variable, a, and increment it.
```python
# create a, and give it an initial value
a = 1
# increment it
a += 1
```
@GPT4-doc-end
"""
code_string = PythonExecutableAnswer.remove_markdown_code(gpt_raw)
return PythonExecutableAnswer(code=code_string, gpt_raw=gpt_raw)
def __str__(self):
return str(self.code)
class TextAnswer(ParsedAnswer):
""" Text, with minimal parsing
"""
def __init__(self, data, gpt_raw:str=None):
self.data = data
self.raw = gpt_raw
@staticmethod
def parser(gpt_raw:str):
data = gpt_raw
return TextAnswer(data, gpt_raw=gpt_raw)
def __str__ (self):
return str(self.data)
# Pass
class YesNoWhy(ParsedAnswer):
""" Yes/No, but with a reason in the end, and/or suggestions.
"""
def __init__(self, final_answer, reason_or_suggestions, gpt_raw:str=None):
self.final_answer = final_answer
self.reason_or_suggestions = reason_or_suggestions
self.raw = gpt_raw
@staticmethod
def parser(gpt_raw:str):
"""
@GPT4-doc-begin
Args:
gpt_raw:
string that contains the tag "[#reason]", where you include the reasoning for
the final decision (yes/no), followed by the tag "[#finalanswer]", followed by the
final answer (yes/no).
For example,
[#reason]
your reasoning for why you will answer yes should be put here.
[#finalanswer]
yes.
@GPT4-doc-end
"""
if "[#reason]" not in gpt_raw:
# print("@@@@@@@@@@@@@@@@@@@@@@@@@@@@@")
# print(gpt_raw)
# print("@@@@@@@@@@@@@@@@@@@@@@@@@@@@@")
raise GPTOutputParseException(f"{gpt_raw} should have [#reason] tag")
if "[#finalanswer]" not in gpt_raw:
# print("@@@@@@@@@@@@@@@@@@@@@@@@@@@@@")
# print(gpt_raw)
# print("@@@@@@@@@@@@@@@@@@@@@@@@@@@@@")
raise GPTOutputParseException(f"{gpt_raw} should have [#finalanswer] tag")
gpt_raw_reasoning = "".join(gpt_raw.split("[#reason]")[1:]).split("[#finalanswer]")[0]
gpt_raw_decision = "".join(gpt_raw.split("[#finalanswer]")[1:])
return YesNoWhy(
final_answer=str(YesNo.parser(gpt_raw_decision)),
reason_or_suggestions=str(TextAnswer.parser(gpt_raw_reasoning)),
gpt_raw=gpt_raw
)
def success(self):
return self.final_answer== "yes"
def __str__(self):
return f"{self.final_answer}: {self.reason_or_suggestions}"
class YesNo(ParsedAnswer):
""" Yes/No
"""
def __init__(self, data, gpt_raw:str=None):
self.data = data
self.raw = gpt_raw
@staticmethod
def remove_punctuation(s):
return re.sub(r'[^\w\s]', '', s)
@staticmethod
def parser(gpt_raw:str):
"""
@GPT4-doc-begin
Args:
gpt_raw: string that contains yes/no answers only, and no other words.
For example,
yes.
@GPT4-doc-end
"""
yesorno = YesNo.remove_punctuation( gpt_raw.lower().strip())
if yesorno== "yes":
data = "yes"
elif yesorno == "no":
data = "no"
else:
# print("@@@@@@@@@@@@@@@@@@@@@@@@@@@@@")
# print(gpt_raw)
# print("@@@@@@@@@@@@@@@@@@@@@@@@@@@@@")
raise GPTOutputParseException(f"{yesorno} cannot be parsed to yes/no")
return YesNo(data, gpt_raw=gpt_raw)
def success(self):
""" used in the context of judging the completion of a task.
"""
return self.data == "yes"
def __str__(self) :
return self.data
class Number(ParsedAnswer):
""" Yes/No
"""
def __init__(self, data, gpt_raw:str=None):
self.data = data
self.raw = gpt_raw
@staticmethod
def remove_punctuation(s):
return re.sub(r'[^\w\s]', '', s)
@staticmethod
def parser(gpt_raw:str) -> "ReadSign":
"""
@GPT4-doc-begin
gpt_raw: string that contains just a simple number.
For example,
90
@GPT4-doc-end
"""
gpt_out = gpt_raw.strip().strip('.').strip(',').lower()
if not gpt_out.isdigit():
raise GPTOutputParseException("output should only contain a number!")
return Number(gpt_out, gpt_raw=gpt_raw)
def success(self):
""" used in the context of judging the completion of a task.
"""
return self.data == "yes"
def __str__(self) :
return self.data |