Spaces:
Sleeping
Sleeping
Create context_parser.py
Browse files- context_parser.py +160 -0
context_parser.py
ADDED
|
@@ -0,0 +1,160 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
from __future__ import annotations
|
| 2 |
+
|
| 3 |
+
import re
|
| 4 |
+
from typing import List, Optional
|
| 5 |
+
|
| 6 |
+
from models import ParsedContext
|
| 7 |
+
|
| 8 |
+
|
| 9 |
+
def extract_section(text: str, start_label: str, end_labels: List[str]) -> str:
|
| 10 |
+
start_idx = text.find(start_label)
|
| 11 |
+
if start_idx == -1:
|
| 12 |
+
return ""
|
| 13 |
+
|
| 14 |
+
start_idx += len(start_label)
|
| 15 |
+
remaining = text[start_idx:]
|
| 16 |
+
|
| 17 |
+
end_positions = []
|
| 18 |
+
for label in end_labels:
|
| 19 |
+
idx = remaining.find(label)
|
| 20 |
+
if idx != -1:
|
| 21 |
+
end_positions.append(idx)
|
| 22 |
+
|
| 23 |
+
if end_positions:
|
| 24 |
+
end_idx = min(end_positions)
|
| 25 |
+
return remaining[:end_idx].strip()
|
| 26 |
+
|
| 27 |
+
return remaining.strip()
|
| 28 |
+
|
| 29 |
+
|
| 30 |
+
def parse_hidden_context(message: str) -> ParsedContext:
|
| 31 |
+
raw = message or ""
|
| 32 |
+
|
| 33 |
+
visible_user_text = ""
|
| 34 |
+
question_text = ""
|
| 35 |
+
options_text = ""
|
| 36 |
+
recent_conversation = ""
|
| 37 |
+
|
| 38 |
+
if "Latest player message:" in raw:
|
| 39 |
+
visible_user_text = extract_section(raw, "Latest player message:", [])
|
| 40 |
+
elif "Player message:" in raw:
|
| 41 |
+
visible_user_text = extract_section(raw, "Player message:", [])
|
| 42 |
+
else:
|
| 43 |
+
visible_user_text = raw.strip()
|
| 44 |
+
|
| 45 |
+
if "Question:" in raw:
|
| 46 |
+
question_text = extract_section(
|
| 47 |
+
raw,
|
| 48 |
+
"Question:",
|
| 49 |
+
["\nOptions:", "\nPlayer balance:", "\nLast outcome:", "\nRecent conversation:", "\nLatest player message:", "\nPlayer message:"]
|
| 50 |
+
)
|
| 51 |
+
|
| 52 |
+
if "Options:" in raw:
|
| 53 |
+
options_text = extract_section(
|
| 54 |
+
raw,
|
| 55 |
+
"Options:",
|
| 56 |
+
["\nPlayer balance:", "\nLast outcome:", "\nRecent conversation:", "\nLatest player message:", "\nPlayer message:"]
|
| 57 |
+
)
|
| 58 |
+
|
| 59 |
+
if "Recent conversation:" in raw:
|
| 60 |
+
recent_conversation = extract_section(
|
| 61 |
+
raw,
|
| 62 |
+
"Recent conversation:",
|
| 63 |
+
["\nLatest player message:", "\nPlayer message:"]
|
| 64 |
+
)
|
| 65 |
+
|
| 66 |
+
combined_parts = []
|
| 67 |
+
if question_text:
|
| 68 |
+
combined_parts.append(question_text.strip())
|
| 69 |
+
if options_text:
|
| 70 |
+
combined_parts.append(options_text.strip())
|
| 71 |
+
|
| 72 |
+
combined_question_block = "\n".join([p for p in combined_parts if p]).strip()
|
| 73 |
+
if not combined_question_block:
|
| 74 |
+
combined_question_block = raw.strip()
|
| 75 |
+
|
| 76 |
+
return ParsedContext(
|
| 77 |
+
raw_message=raw,
|
| 78 |
+
visible_user_text=visible_user_text.strip(),
|
| 79 |
+
full_context_text=raw.strip(),
|
| 80 |
+
question_text=question_text.strip(),
|
| 81 |
+
options_text=options_text.strip(),
|
| 82 |
+
combined_question_block=combined_question_block.strip(),
|
| 83 |
+
recent_conversation=recent_conversation.strip(),
|
| 84 |
+
)
|
| 85 |
+
|
| 86 |
+
|
| 87 |
+
def detect_help_mode(user_text: str, supplied: Optional[str]) -> str:
|
| 88 |
+
lower = (user_text or "").lower().strip()
|
| 89 |
+
|
| 90 |
+
if supplied in {"hint", "walkthrough", "answer"}:
|
| 91 |
+
if lower not in {"hint", "walkthrough", "answer"}:
|
| 92 |
+
if "hint" in lower:
|
| 93 |
+
return "hint"
|
| 94 |
+
if "walkthrough" in lower or "step by step" in lower or "work through" in lower:
|
| 95 |
+
return "walkthrough"
|
| 96 |
+
if "explain" in lower and "why" in lower:
|
| 97 |
+
return "walkthrough"
|
| 98 |
+
return supplied
|
| 99 |
+
return supplied
|
| 100 |
+
|
| 101 |
+
if lower in {"hint", "just a hint", "hint please", "small hint"}:
|
| 102 |
+
return "hint"
|
| 103 |
+
|
| 104 |
+
if "hint" in lower:
|
| 105 |
+
return "hint"
|
| 106 |
+
|
| 107 |
+
if (
|
| 108 |
+
"walkthrough" in lower
|
| 109 |
+
or "step by step" in lower
|
| 110 |
+
or "work through" in lower
|
| 111 |
+
or "explain step by step" in lower
|
| 112 |
+
):
|
| 113 |
+
return "walkthrough"
|
| 114 |
+
|
| 115 |
+
if "why" in lower or "explain" in lower:
|
| 116 |
+
return "walkthrough"
|
| 117 |
+
|
| 118 |
+
return "answer"
|
| 119 |
+
|
| 120 |
+
|
| 121 |
+
def user_is_referring_to_existing_question(user_text: str, question_text: str) -> bool:
|
| 122 |
+
lower = (user_text or "").lower().strip()
|
| 123 |
+
|
| 124 |
+
if not question_text:
|
| 125 |
+
return False
|
| 126 |
+
|
| 127 |
+
short_refs = {
|
| 128 |
+
"help", "hint", "why", "why?", "explain", "explain it", "explain this",
|
| 129 |
+
"answer", "what's the answer", "what is the answer", "is it a", "is it b",
|
| 130 |
+
"is it c", "is it d", "is it e", "which one", "which option",
|
| 131 |
+
"what do i do first", "what do i do", "how do i start", "are you sure",
|
| 132 |
+
"can you help", "help me", "i don't get it", "i dont get it",
|
| 133 |
+
"why is it c", "why is it b", "why is it a", "why is it d", "why is it e"
|
| 134 |
+
}
|
| 135 |
+
|
| 136 |
+
if lower in short_refs:
|
| 137 |
+
return True
|
| 138 |
+
|
| 139 |
+
if len(lower) <= 25 and any(phrase in lower for phrase in [
|
| 140 |
+
"is it ", "why", "hint", "explain", "answer", "help", "first step"
|
| 141 |
+
]):
|
| 142 |
+
return True
|
| 143 |
+
|
| 144 |
+
return False
|
| 145 |
+
|
| 146 |
+
|
| 147 |
+
def mentions_choice_letter(user_text: str) -> Optional[str]:
|
| 148 |
+
lower = (user_text or "").lower().strip()
|
| 149 |
+
m = re.search(r"\b(?:is it|answer is|it's|its|option|choice)\s*([a-e])\b", lower)
|
| 150 |
+
if m:
|
| 151 |
+
return m.group(1).upper()
|
| 152 |
+
|
| 153 |
+
m = re.search(r"\bwhy is it\s*([a-e])\b", lower)
|
| 154 |
+
if m:
|
| 155 |
+
return m.group(1).upper()
|
| 156 |
+
|
| 157 |
+
if lower in {"a", "b", "c", "d", "e"}:
|
| 158 |
+
return lower.upper()
|
| 159 |
+
|
| 160 |
+
return None
|