mirror of
https://github.com/InternLM/InternBootcamp.git
synced 2026-04-25 17:10:49 +00:00
init-commit
This commit is contained in:
lilinyang
commit
18a552597a
3461 changed files with 1150579 additions and 0 deletions
228
internbootcamp/bootcamp/campsite/campsite_default.py
Executable file
228
internbootcamp/bootcamp/campsite/campsite_default.py
Executable file
|
|
@ -0,0 +1,228 @@
|
|||
import re
|
||||
import ast
|
||||
import json
|
||||
import sys
|
||||
import random
|
||||
sys.path.append('./')
|
||||
from internbootcamp.bootcamp.base import Basebootcamp
|
||||
# from environments import CampsiteEnvironment
|
||||
from internbootcamp.libs.campsite.campsite_generator import generate_campsite
|
||||
from internbootcamp.libs.campsite.campsite_validor import validate_campsite_solution
|
||||
|
||||
class Campsitebootcamp(Basebootcamp):
|
||||
|
||||
def __init__(self, size:tuple = (8,8), expect_camp_number:int = 8, random_rate:float = 0.1 ):
|
||||
super().__init__()
|
||||
self.size = size
|
||||
self.expect_camp_number = expect_camp_number
|
||||
self.random_rate = random_rate
|
||||
|
||||
# self.env = CampsiteEnvironment()
|
||||
|
||||
|
||||
def generator(self, size:tuple = (8,8), expect_camp_number:int = 8, random_rate:float = 0.1 , seed:int = None):
|
||||
if size[0] < 2 and size[1] < 2:
|
||||
raise ValueError
|
||||
self.grid, self.row_constraints, self.col_constraints, self.refer_ans = generate_campsite(size[0], size[1], expect_camp_number, seed=seed, random_rate=random_rate)
|
||||
return self.grid, self.row_constraints, self.col_constraints
|
||||
|
||||
def case_generator(self):
|
||||
grid, row_constraints, col_constraints = self.generator(self.size,self.expect_camp_number, self.random_rate )
|
||||
|
||||
# ans, output, all_ans= self.env.solver()
|
||||
|
||||
# self.env.reset()
|
||||
|
||||
return {
|
||||
'input_grid': grid,
|
||||
'row_constraints': row_constraints,
|
||||
'col_constraints': col_constraints
|
||||
}
|
||||
|
||||
def prompt_func(self, identity) -> str:
|
||||
"""
|
||||
Process the input_data and return the processed prompt.
|
||||
|
||||
Args:
|
||||
question_ori: The question to be processed.
|
||||
|
||||
Returns:
|
||||
str: The processed prompt.
|
||||
"""
|
||||
|
||||
statements = [f"""You are an intelligent assistant specializing in solving custom puzzle problems. Below is a specific rule defined for a custom puzzle. Your task is to apply this rule accurately to the provided question.
|
||||
|
||||
### Instructions:
|
||||
|
||||
1. Thoroughly understand the rule provided. If needed, break down the rule into simpler components or steps.
|
||||
2. Apply the rule carefully to address the question presented.
|
||||
3. Verify your answer to ensure it aligns with the rule and the context of the puzzle.
|
||||
|
||||
### Campsite Puzzle Rule:
|
||||
|
||||
1.The game is played on an n*m grid with trees at some locations in the grid.
|
||||
2.To place tents on the grid.
|
||||
3.Each tent must be orthogonally adjacent to a tree (i.e., above, below, to the left, or to the right of the tree).
|
||||
4.Tents cannot be orthogonally or diagonally adjacent to other tents.
|
||||
5.Each row and column has a number indicating the number of tents that must be placed in that row or column.
|
||||
6.Each puzzle has and has only one unique solution.
|
||||
7.The puzzle is given by a matrix in the form of T, which represents the position of the tree, and X, which represents the spaces,To the right and below the matrix are numerical constraints, and you need to replace X with C (for tents) for some spaces, and the answer is a matrix.
|
||||
|
||||
### Question
|
||||
|
||||
Now the gird of the Campsite is:\n{identity['input_grid']}
|
||||
The row constraints is {identity['row_constraints']} and the col constraints is {identity['col_constraints']}.
|
||||
""",
|
||||
f"""The Campsite is a puzzle game. The rule of Campsite is:
|
||||
1.The game is played on an n*m grid with trees at some locations in the grid.
|
||||
2.To place tents on the grid.
|
||||
3.Each tent must be orthogonally adjacent to a tree (i.e., above, below, to the left, or to the right of the tree).
|
||||
4.Tents cannot be orthogonally or diagonally adjacent to other tents.
|
||||
5.Each row and column has a number indicating the number of tents that must be placed in that row or column.
|
||||
6.Each puzzle has and has only one unique solution.
|
||||
7.The puzzle is given by a matrix in the form of T, which represents the position of the tree, and X, which represents the spaces,To the right and below the matrix are numerical constraints, and you need to replace X with C (for tents) for some spaces, and the answer is a matrix.
|
||||
|
||||
Now the gird of the Campsite is:\n{identity['input_grid']}
|
||||
The row constraints is {identity['row_constraints']} and the col constraints is {identity['col_constraints']}.
|
||||
"""
|
||||
]
|
||||
|
||||
instruction_following = """
|
||||
Let's think step by step and output the final answer with an example json formatting
|
||||
Final-answer: ```json
|
||||
[['T', 'X', 'X', 'X'], ['X', 'T', 'X', 'X'], ['T', 'X', 'X', 'T'], ['X', 'X', 'X', 'X']]
|
||||
```
|
||||
"""
|
||||
|
||||
return statements[random.randint(0,len(statements)-1)] + instruction_following
|
||||
|
||||
@staticmethod
|
||||
def extract_output(output):
|
||||
"""
|
||||
Extract the output from the solution.
|
||||
|
||||
Args:
|
||||
output: Model output to be processed.
|
||||
|
||||
Returns:
|
||||
The processed output.
|
||||
"""
|
||||
pattern = pattern = r'```json\s*([\s\S]*?)\s*```'
|
||||
matches = re.findall(pattern, output)
|
||||
# Final-answer:{'A': [(3, 1)],
|
||||
# 'B': [(1, 4)],
|
||||
# 'C': [(2, 3)],
|
||||
# 'D': [(5, 2)],
|
||||
# 'E': [(4, 5)]}
|
||||
if matches:
|
||||
# 获取 JSON 字符串
|
||||
json_str = matches[-1]
|
||||
# print('match?', json_str)
|
||||
# print('solution generated? first lines', output[:200])
|
||||
# print('solution generated? last lines', output[-200:])
|
||||
# 替换单引号为双引号，将元组表示改为列表表示
|
||||
json_str = json_str.replace("'", '"').replace("(", "[").replace(")", "]")
|
||||
try:
|
||||
# 解析 JSON 字符串为 Python 字典
|
||||
result_dict = json.loads(json_str)
|
||||
return result_dict
|
||||
except json.JSONDecodeError as e:
|
||||
return json_str
|
||||
else:
|
||||
return None
|
||||
|
||||
|
||||
@staticmethod
|
||||
def check_solution(parsed_question: dict, parsed_response: dict) -> bool:
|
||||
original_grid = parsed_question['input_grid']
|
||||
expected_rows = parsed_question['row_constraints']
|
||||
expected_cols = parsed_question['col_constraints']
|
||||
solution = parsed_response
|
||||
|
||||
n, m = len(original_grid), len(original_grid[0]) if original_grid else 0
|
||||
|
||||
# Check dimensions
|
||||
if len(solution) != n or any(len(row) != m for row in solution):
|
||||
return False
|
||||
|
||||
tents = []
|
||||
# Check T positions and collect tents
|
||||
for i in range(n):
|
||||
for j in range(m):
|
||||
orig = original_grid[i][j]
|
||||
sol = solution[i][j]
|
||||
if orig == 'T':
|
||||
if sol != 'T':
|
||||
return False
|
||||
else:
|
||||
if sol not in ('X', 'C'):
|
||||
return False
|
||||
if sol == 'C':
|
||||
tents.append((i, j))
|
||||
|
||||
# Check tents adjacency to trees and other tents
|
||||
directions_ortho = [(-1, 0), (1, 0), (0, -1), (0, 1)]
|
||||
for i, j in tents:
|
||||
has_tree = False
|
||||
for dx, dy in directions_ortho:
|
||||
x, y = i + dx, j + dy
|
||||
if 0 <= x < n and 0 <= y < m and solution[x][y] == 'T':
|
||||
has_tree = True
|
||||
break
|
||||
if not has_tree:
|
||||
return False
|
||||
|
||||
for dx in (-1, 0, 1):
|
||||
for dy in (-1, 0, 1):
|
||||
if dx == 0 and dy == 0:
|
||||
continue
|
||||
x, y = i + dx, j + dy
|
||||
if 0 <= x < n and 0 <= y < m and (x, y) in tents:
|
||||
return False
|
||||
|
||||
# Check row constraints
|
||||
for i in range(n):
|
||||
if sum(1 for cell in solution[i] if cell == 'C') != expected_rows[i]:
|
||||
return False
|
||||
|
||||
# Check column constraints
|
||||
for j in range(m):
|
||||
if sum(1 for i in range(n) if solution[i][j] == 'C') != expected_cols[j]:
|
||||
return False
|
||||
|
||||
return True
|
||||
|
||||
@classmethod
|
||||
def _verify_correction(cls, solution, identity):
|
||||
input_grid = identity['input_grid']
|
||||
row_constraints = identity['row_constraints']
|
||||
col_constraints = identity['col_constraints']
|
||||
input_grid = json.loads(input_grid) if type(input_grid) == str else input_grid
|
||||
row_constraints = json.loads(row_constraints) if type(row_constraints) == str else row_constraints
|
||||
col_constraints = json.loads(col_constraints) if type(col_constraints) == str else col_constraints
|
||||
is_valid, msg = validate_campsite_solution(puzzle=input_grid, row_constraints=row_constraints, col_constraints=col_constraints, solution=solution)
|
||||
return is_valid
|
||||
|
||||
|
||||
if __name__ == "__main__":
|
||||
|
||||
# test verify
|
||||
identity = {
|
||||
"input_grid": [
|
||||
["T", "X", "X", "X"],
|
||||
["X", "T", "X", "X"],
|
||||
["T", "X", "X", "T"],
|
||||
["X", "X", "X", "X"]
|
||||
],
|
||||
"row_constraints": [0, 2, 0, 2],
|
||||
"col_constraints": [2, 0, 1, 1]
|
||||
}
|
||||
|
||||
solution = """```json
|
||||
[
|
||||
["T", "X", "X", "X"],
|
||||
["C", "T", "C", "X"],
|
||||
["T", "X", "X", "T"],
|
||||
["C", "X", "X", "C"]
|
||||
]```"""
|
||||
print(Campsitebootcamp.verify_score(model_output=solution, identity=identity))
|
||||
Loading…
Add table
Add a link
Reference in a new issue