1#!/usr/bin/env python3
2# -*- coding: utf-8 -*-
3"""
4CIP Design Core - BM25 search engine for Corporate Identity Program design guidelines
5"""
6 
7import csv
8import re
9from pathlib import Path
10from math import log
11from collections import defaultdict
12 
13# ============ CONFIGURATION ============
14DATA_DIR = Path(__file__).parent.parent.parent / "data" / "cip"
15MAX_RESULTS = 3
16 
17CSV_CONFIG = {
18    "deliverable": {
19        "file": "deliverables.csv",
20        "search_cols": ["Deliverable", "Category", "Keywords", "Description", "Mockup Context"],
21        "output_cols": ["Deliverable", "Category", "Keywords", "Description", "Dimensions", "File Format", "Logo Placement", "Color Usage", "Typography Notes", "Mockup Context", "Best Practices", "Avoid"]
22    },
23    "style": {
24        "file": "styles.csv",
25        "search_cols": ["Style Name", "Category", "Keywords", "Description", "Mood"],
26        "output_cols": ["Style Name", "Category", "Keywords", "Description", "Primary Colors", "Secondary Colors", "Typography", "Materials", "Finishes", "Mood", "Best For", "Avoid For"]
27    },
28    "industry": {
29        "file": "industries.csv",
30        "search_cols": ["Industry", "Keywords", "CIP Style", "Mood"],
31        "output_cols": ["Industry", "Keywords", "CIP Style", "Primary Colors", "Secondary Colors", "Typography", "Key Deliverables", "Mood", "Best Practices", "Avoid"]
32    },
33    "mockup": {
34        "file": "mockup-contexts.csv",
35        "search_cols": ["Context Name", "Category", "Keywords", "Scene Description"],
36        "output_cols": ["Context Name", "Category", "Keywords", "Scene Description", "Lighting", "Environment", "Props", "Camera Angle", "Background", "Style Notes", "Best For", "Prompt Modifiers"]
37    }
38}
39 
40 
41# ============ BM25 IMPLEMENTATION ============
42class BM25:
43    """BM25 ranking algorithm for text search"""
44 
45    def __init__(self, k1=1.5, b=0.75):
46        self.k1 = k1
47        self.b = b
48        self.corpus = []
49        self.doc_lengths = []
50        self.avgdl = 0
51        self.idf = {}
52        self.doc_freqs = defaultdict(int)
53        self.N = 0
54 
55    def tokenize(self, text):
56        """Lowercase, split, remove punctuation, filter short words"""
57        text = re.sub(r'[^\w\s]', ' ', str(text).lower())
58        return [w for w in text.split() if len(w) > 2]
59 
60    def fit(self, documents):
61        """Build BM25 index from documents"""
62        self.corpus = [self.tokenize(doc) for doc in documents]
63        self.N = len(self.corpus)
64        if self.N == 0:
65            return
66        self.doc_lengths = [len(doc) for doc in self.corpus]
67        self.avgdl = sum(self.doc_lengths) / self.N
68 
69        for doc in self.corpus:
70            seen = set()
71            for word in doc:
72                if word not in seen:
73                    self.doc_freqs[word] += 1
74                    seen.add(word)
75 
76        for word, freq in self.doc_freqs.items():
77            self.idf[word] = log((self.N - freq + 0.5) / (freq + 0.5) + 1)
78 
79    def score(self, query):
80        """Score all documents against query"""
81        query_tokens = self.tokenize(query)
82        scores = []
83 
84        for idx, doc in enumerate(self.corpus):
85            score = 0
86            doc_len = self.doc_lengths[idx]
87            term_freqs = defaultdict(int)
88            for word in doc:
89                term_freqs[word] += 1
90 
91            for token in query_tokens:
92                if token in self.idf:
93                    tf = term_freqs[token]
94                    idf = self.idf[token]
95                    numerator = tf * (self.k1 + 1)
96                    denominator = tf + self.k1 * (1 - self.b + self.b * doc_len / self.avgdl)
97                    score += idf * numerator / denominator
98 
99            scores.append((idx, score))
100 
101        return sorted(scores, key=lambda x: x[1], reverse=True)
102 
103 
104# ============ SEARCH FUNCTIONS ============
105def _load_csv(filepath):
106    """Load CSV and return list of dicts"""
107    with open(filepath, 'r', encoding='utf-8') as f:
108        return list(csv.DictReader(f))
109 
110 
111def _search_csv(filepath, search_cols, output_cols, query, max_results):
112    """Core search function using BM25"""
113    if not filepath.exists():
114        return []
115 
116    data = _load_csv(filepath)
117 
118    # Build documents from search columns
119    documents = [" ".join(str(row.get(col, "")) for col in search_cols) for row in data]
120 
121    # BM25 search
122    bm25 = BM25()
123    bm25.fit(documents)
124    ranked = bm25.score(query)
125 
126    # Get top results with score > 0
127    results = []
128    for idx, score in ranked[:max_results]:
129        if score > 0:
130            row = data[idx]
131            results.append({col: row.get(col, "") for col in output_cols if col in row})
132 
133    return results
134 
135 
136def detect_domain(query):
137    """Auto-detect the most relevant domain from query"""
138    query_lower = query.lower()
139 
140    domain_keywords = {
141        "deliverable": ["card", "letterhead", "envelope", "folder", "shirt", "cap", "badge", "signage", "vehicle", "car", "van", "stationery", "uniform", "merchandise", "packaging", "banner", "booth"],
142        "style": ["style", "minimal", "modern", "luxury", "vintage", "industrial", "elegant", "bold", "corporate", "organic", "playful"],
143        "industry": ["tech", "finance", "legal", "healthcare", "hospitality", "food", "fashion", "retail", "construction", "logistics"],
144        "mockup": ["mockup", "scene", "context", "photo", "shot", "lighting", "background", "studio", "lifestyle"]
145    }
146 
147    scores = {domain: sum(1 for kw in keywords if kw in query_lower) for domain, keywords in domain_keywords.items()}
148    best = max(scores, key=scores.get)
149    return best if scores[best] > 0 else "deliverable"
150 
151 
152def search(query, domain=None, max_results=MAX_RESULTS):
153    """Main search function with auto-domain detection"""
154    if domain is None:
155        domain = detect_domain(query)
156 
157    config = CSV_CONFIG.get(domain, CSV_CONFIG["deliverable"])
158    filepath = DATA_DIR / config["file"]
159 
160    if not filepath.exists():
161        return {"error": f"File not found: {filepath}", "domain": domain}
162 
163    results = _search_csv(filepath, config["search_cols"], config["output_cols"], query, max_results)
164 
165    return {
166        "domain": domain,
167        "query": query,
168        "file": config["file"],
169        "count": len(results),
170        "results": results
171    }
172 
173 
174def search_all(query, max_results=2):
175    """Search across all domains and combine results"""
176    all_results = {}
177    for domain in CSV_CONFIG.keys():
178        result = search(query, domain, max_results)
179        if result.get("results"):
180            all_results[domain] = result["results"]
181    return all_results
182 
183 
184def get_cip_brief(brand_name, industry_query, style_query=None):
185    """Generate a comprehensive CIP brief for a brand"""
186    # Search industry
187    industry_results = search(industry_query, "industry", 1)
188    industry = industry_results.get("results", [{}])[0] if industry_results.get("results") else {}
189 
190    # Search style (use industry style if not specified)
191    style_query = style_query or industry.get("CIP Style", "corporate minimal")
192    style_results = search(style_query, "style", 1)
193    style = style_results.get("results", [{}])[0] if style_results.get("results") else {}
194 
195    # Get recommended deliverables for the industry
196    key_deliverables = industry.get("Key Deliverables", "").split()
197    deliverable_results = []
198    for d in key_deliverables[:5]:
199        result = search(d, "deliverable", 1)
200        if result.get("results"):
201            deliverable_results.append(result["results"][0])
202 
203    return {
204        "brand_name": brand_name,
205        "industry": industry,
206        "style": style,
207        "recommended_deliverables": deliverable_results,
208        "color_system": {
209            "primary": style.get("Primary Colors", industry.get("Primary Colors", "")),
210            "secondary": style.get("Secondary Colors", industry.get("Secondary Colors", ""))
211        },
212        "typography": style.get("Typography", industry.get("Typography", "")),
213        "materials": style.get("Materials", ""),
214        "finishes": style.get("Finishes", "")
215    }
216