File size: 5,350 Bytes
06ba7ea | 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 | import json
import re
from typing import Any, Dict, Optional, Iterable
def try_parse_tool_call(text:str) -> Optional[Dict[str, Any]]:
"""
Return dict if text is a valid tool call JSON, otherwise return None
"""
try:
obj = parse_json_dict(text)
except:
return None
if obj.get("action") != "call_tool":
return None
if "tool" not in obj:
return None
args = obj.get("arguments", {})
if args is not None and not isinstance(args, dict):
return None
return obj
# Support ```json ... ``` and ```jsonc ... ``` (can remove jsonc if needed)
_CODE_FENCE_RE = re.compile(
r"```(?:json|jsonc)\s*(.*?)\s*```",
flags=re.IGNORECASE | re.DOTALL,
)
def _strip_trailing_commas_once(s: str) -> str:
"""
Remove trailing commas before '}' or ']' in JSON text (single pass).
Note: Skips content inside strings, won't remove commas within strings.
"""
out = []
in_str = False
escape = False
i = 0
n = len(s)
while i < n:
c = s[i]
if in_str:
out.append(c)
if escape:
escape = False
elif c == "\\":
escape = True
elif c == '"':
in_str = False
i += 1
continue
# not in string
if c == '"':
in_str = True
out.append(c)
i += 1
continue
if c == ",":
# look ahead to next non-whitespace
j = i + 1
while j < n and s[j] in " \t\r\n":
j += 1
if j < n and s[j] in "}]":
# drop this comma
i += 1
continue
out.append(c)
i += 1
return "".join(out)
def _strip_trailing_commas(s: str, max_passes: int = 10) -> str:
"""
Remove extra commas before '}' or ']' in JSON text (single pass).
Note: String content is skipped, so commas inside strings won't be removed.
"""
for _ in range(max_passes):
s2 = _strip_trailing_commas_once(s)
if s2 == s:
return s2
s = s2
return s # best effort
def _extract_balanced_object(text: str, start: int) -> Optional[str]:
"""
Extract a balanced JSON object substring {...} starting from text[start] == '{'.
Correctly skips braces within strings.
"""
depth = 0
in_str = False
escape = False
for i in range(start, len(text)):
c = text[i]
if in_str:
if escape:
escape = False
elif c == "\\":
escape = True
elif c == '"':
in_str = False
continue
if c == '"':
in_str = True
continue
if c == "{":
depth += 1
elif c == "}":
depth -= 1
if depth == 0:
return text[start : i + 1]
return None
def _iter_fenced_json_blocks(text: str) -> Iterable[str]:
for m in _CODE_FENCE_RE.finditer(text):
block = m.group(1)
if block is not None:
yield block.strip()
def _iter_object_candidates(text: str) -> Iterable[str]:
"""
Enumerate all possible {...} substrings in arbitrary text (in order of appearance).
"""
for idx, ch in enumerate(text):
if ch == "{":
cand = _extract_balanced_object(text, idx)
if cand:
yield cand
def parse_json_dict(text: str) -> Dict[str, Any]:
"""
Parse a JSON object (dict) from arbitrary text.
Supports:
1) Markdown fenced JSON code blocks: ```json ... ```
2) JSON surrounded by extra text
3) Removing trailing commas before '}' or ']'
Args:
text: Input string to parse
Returns:
Parsed dictionary
Raises:
ValueError: Cannot find a valid JSON dict to parse
TypeError: Input text is not a string
"""
if not isinstance(text, str):
raise TypeError(f"text must be str, got {type(text).__name__}")
# Try fenced block first, then try the entire text
search_spaces = list(_iter_fenced_json_blocks(text))
search_spaces.append(text)
last_err: Optional[Exception] = None
for space in search_spaces:
# If starts with '{', try to extract a balanced object from the beginning first (to avoid trailing noise)
candidates = []
stripped = space.lstrip().lstrip("\ufeff") # 顺便去 BOM
if stripped.startswith("{"):
first = _extract_balanced_object(stripped, 0)
if first:
candidates.append(first)
# Also try objects appearing at any position in the text
candidates.extend(_iter_object_candidates(space))
# Deduplicate (avoid retrying same substrings)
seen = set()
for cand in candidates:
if cand in seen:
continue
seen.add(cand)
cleaned = _strip_trailing_commas(cand).strip()
try:
obj = json.loads(cleaned)
if isinstance(obj, dict):
return obj
except Exception as e:
last_err = e
continue
raise ValueError("No valid JSON object (dict) found in input") from last_err |