File size: 10,030 Bytes
41fe3fc
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
"""Small shared helpers: HTML escaping, metrics, history windowing, LLM factory, sanitizers."""
import asyncio
import html
import json
import logging
import os
import re
import time
import uuid
from datetime import datetime
from typing import Any, Dict, List, Literal, Optional, Tuple
from urllib.parse import quote

import httpx
from pydantic import BaseModel, ConfigDict, Field

from src.config import get_settings, LIBBEE_VERSION
from src.services.staff_service import (
    STAFF_DIRECTORY,
    match_staff_name,
    match_staff_role,
    should_attempt_staff_lookup,
    staff_name_answer,
    staff_role_answer,
)

from src.agentcore.models import ChatMessage, SearchContextPayload
from src.agentcore.constants import (
    CURRENT_YEAR,
    HISTORY_WINDOW,
    _LEGACY_SYSTEM_PATTERNS,
    _RESOURCE_TYPE_NOISE,
)

logger = logging.getLogger(__name__)

def _get_runtime_config() -> dict:
    try:
        from src.services.runtime_store import JsonRuntimeStore
        settings = get_settings()
        store = JsonRuntimeStore(settings.config_path, default={
            "max_results": 5,
            "maintenance_mode": False,
            "welcome_message": "",
            "custom_instructions": "",
            "announcement": "",
            "maintenance_message": "",
        })
        return store.load()
    except Exception:
        return {}


def _escape(text: str) -> str:
    return html.escape(text or "")


def _normalize_whitespace(text: str) -> str:
    return re.sub(r"\s+", " ", (text or "").strip())


def _title_case_topic(topic: str) -> str:
    return _normalize_whitespace(topic).strip().strip(".?")


def _strip_resource_noise(text: str) -> str:
    cleaned = _RESOURCE_TYPE_NOISE.sub(" ", text or "")
    cleaned = re.sub(r"\s+(AND|OR)\s+(AND|OR)\s+", " AND ", cleaned, flags=re.IGNORECASE)
    cleaned = re.sub(r"^\s*(AND|OR)\s+", "", cleaned, flags=re.IGNORECASE)
    cleaned = re.sub(r"\s+(AND|OR)\s*$", "", cleaned, flags=re.IGNORECASE)
    return _normalize_whitespace(cleaned)


def _safe_metrics_increment(key: str) -> None:
    try:
        from app import get_metrics_service
        get_metrics_service().incr(key)
    except Exception:
        return


def _safe_metrics_bucket(bucket: str, key: str) -> None:
    try:
        from app import get_metrics_service
        get_metrics_service().incr_bucket(bucket, key)
    except Exception:
        return


def _build_history_messages(history: List[ChatMessage]) -> List[dict]:
    msgs = []
    for m in history[-HISTORY_WINDOW:]:
        if m.role in ("user", "assistant") and m.content:
            msgs.append({"role": m.role, "content": m.content[:300]})
    return msgs


def _get_llm(model: str, temperature: float, max_tokens: int):
    settings = get_settings()
    if model == "claude" and settings.anthropic_api_key:
        from langchain_anthropic import ChatAnthropic
        return ChatAnthropic(
            model="claude-haiku-4-5-20251001",
            temperature=temperature,
            max_tokens=max_tokens,
            anthropic_api_key=settings.anthropic_api_key,
        )
    from langchain_openai import ChatOpenAI
    return ChatOpenAI(
        model="gpt-4o-mini",
        temperature=temperature,
        max_tokens=max_tokens,
        openai_api_key=settings.openai_api_key,
    )


def _shared_build_primo_boolean_query(topic: str) -> str:
    clean = _strip_resource_noise(topic)
    if not clean:
        clean = topic
    clean = _normalize_whitespace(clean)
    words = clean.split()
    if len(words) <= 4:
        return f'"{clean}"'
    _BOOL_STOP = re.compile(
        r"\b(of|on|in|the|a|an|and|or|for|to|with|by|from|at|is|are|was|were|"
        r"be|been|have|has|had|do|does|did|will|would|could|should|may|its|"
        r"this|that|these|those|about|impact|role|effect|use|analysis|review|"
        r"what|how|why|when|where|which|between|within|across|among|using|"
        r"based|related|towards|toward|during|after|before|over|under)\b",
        re.IGNORECASE,
    )
    parts = _BOOL_STOP.split(clean)
    concepts = [_normalize_whitespace(p) for p in parts if _normalize_whitespace(p) and len(_normalize_whitespace(p)) > 2]
    if not concepts:
        return f'"{clean}"'
    if len(concepts) == 1:
        return f'"{concepts[0]}"'
    quoted = [f'"{c}"' if ' ' in c else c for c in concepts[:4]]
    return " AND ".join(quoted)


def _make_primo_boolean_query(context: SearchContextPayload) -> str:
    topic = _title_case_topic(context.display_topic or context.topic)
    topic = _strip_resource_noise(topic) or (context.topic or "library search")
    return _shared_build_primo_boolean_query(topic)


def _shared_build_primo_discovery_url(
    boolean_query: str,
    resource_type: str = "articles",
    peer_reviewed: bool = False,
    open_access: bool = False,
    year_from: Optional[str] = None,
    year_to: Optional[str] = None,
) -> str:
    base = (
        "https://khalifa.primo.exlibrisgroup.com/discovery/search"
        f"?vid=971KUOSTAR_INST:KU&tab=Everything&scope=MyInst_and_CI"
        f"&query=any,contains,{quote(boolean_query)}"
        f"&lang=en&search_scope=MyInst_and_CI&sortby=rank&mode=advanced"
    )
    facets = []
    if resource_type == "articles":
        facets.append("facet_rtype,include,articles")
    elif resource_type == "books":
        facets.append("facet_rtype,include,books")
    if peer_reviewed:
        facets.append("facet_tlevel,include,peer_reviewed")
    if open_access:
        facets.append("facet_tlevel,include,online_resources")
    if year_from or year_to:
        yf = year_from or "0001"
        yt = year_to or "9999"
        facets.append(f"facet_searchcreationdate,include,{yf}|,|{yt}")
    for facet in facets:
        base += f"&multiFacets={quote(facet)}"
    return base


def _shared_build_pubmed_url(
    topic: str,
    year_from: Optional[str] = None,
    year_to: Optional[str] = None,
    peer_reviewed: bool = False,
) -> str:
    clean = _strip_resource_noise(topic)
    term = clean or topic
    if peer_reviewed:
        term = f"({term}) AND Journal Article[pt]"
    url = f"https://pubmed.ncbi.nlm.nih.gov/?term={quote(term)}"
    if year_from or year_to:
        yf = year_from or "1900"
        yt = year_to or str(CURRENT_YEAR)
        url += f"&filter=datesearch.y_{yf}-{yt}"
    return url


def _primo_clean_url(context: SearchContextPayload) -> str:
    boolean_query = context.primo_boolean_query or _make_primo_boolean_query(context)
    return _shared_build_primo_discovery_url(
        boolean_query,
        resource_type=context.resource_type,
        peer_reviewed=context.peer_reviewed,
        open_access=context.open_access,
        year_from=context.year_from,
        year_to=context.year_to,
    )


async def _grammar_refine_query(text: str, model: str) -> str:
    settings = get_settings()
    if not settings.openai_api_key and not settings.anthropic_api_key:
        return _normalize_whitespace(text)
    try:
        llm = _get_llm(model, temperature=0, max_tokens=60)
        response = await llm.ainvoke([
            {"role": "system", "content": "Rewrite the user's search question in clear grammatical English. Keep the meaning exactly the same. Return one sentence only."},
            {"role": "user", "content": text},
        ])
        refined = _normalize_whitespace(response.content)
        return refined or _normalize_whitespace(text)
    except Exception:
        return _normalize_whitespace(text)


def _light_strip_retrieval_boilerplate(text: str) -> str:
    cleaned = re.sub(r"^\s*(please\s+)?(?:can you|could you|would you)\s+", "", (text or "").strip(), flags=re.IGNORECASE)
    cleaned = re.sub(r"^\s*(please\s+)?help me\s+", "", cleaned, flags=re.IGNORECASE)
    cleaned = re.sub(r"^\s*please\s+", "", cleaned, flags=re.IGNORECASE)
    cleaned = re.sub(r"\s+(please|thanks|thank you|asap)$", "", cleaned, flags=re.IGNORECASE)
    cleaned = re.sub(
        r"^\s*(find|search for|search|look for|get me|show me|give me|fetch|retrieve|"
        r"research on|articles on|papers on|literature on|studies on|"
        r"tell me about|i need|i want|can you find|help me find|"
        r"i am looking for|i'm looking for|i need articles on|"
        r"i want articles on|i need papers on|i want papers on)\s+",
        "", cleaned, flags=re.IGNORECASE
    )
    cleaned = re.sub(
        r"^\s*(research|articles?|papers?|books?|literature|studies|study|"
        r"journals?|publications?|resources?)\s+(on|about|for|regarding|into)\s+",
        "", cleaned, flags=re.IGNORECASE
    )
    cleaned = re.sub(
        r"^\s*(on|about|for|regarding|concerning|into|around|of|in|the)\s+",
        "", cleaned, flags=re.IGNORECASE
    )
    return re.sub(r"\s+", " ", cleaned).strip()


def _sanitize_llm_response(text: str) -> str:
    if not text:
        return text
    for pattern, replacement in _LEGACY_SYSTEM_PATTERNS:
        text = pattern.sub(replacement, text)
    return text


def _sanitize_boolean_for_primo(boolean: str) -> str:
    if not boolean:
        return boolean
    boolean = re.sub(r"'([^']+)'", r'"\1"', boolean)
    boolean = re.sub(
        r"\(\s*(?:(?:19|20)\d{2}\s*(?:OR\s*(?:19|20)\d{2}\s*)*)\)",
        "", boolean, flags=re.IGNORECASE,
    )
    boolean = re.sub(r"^\s*(AND|OR)\s*", "", boolean, flags=re.IGNORECASE)
    boolean = re.sub(r"\s*(AND|OR)\s*$", "", boolean, flags=re.IGNORECASE)
    boolean = re.sub(r"\b(AND|OR)\s+(AND|OR)\b", r"\1", boolean, flags=re.IGNORECASE)
    return re.sub(r"\s+", " ", boolean).strip()


def _clean_database_keywords(boolean_query: str) -> str:
    return re.sub(r"\s+", " ",
                  re.sub(r"\b(AND|OR|NOT)\b|[()\"]", " ",
                         boolean_query, flags=re.IGNORECASE)).strip()


def _find_staff_by_token(token: str) -> Optional[dict]:
    token = (token or "").lower()
    for staff in STAFF_DIRECTORY:
        hay = (staff.get("full_name", "") + " " + staff.get("role", "")).lower()
        if token in hay:
            return staff
    return None