|
| 1 | +package main |
| 2 | + |
| 3 | +import ( |
| 4 | + "bufio" |
| 5 | + "encoding/json" |
| 6 | + "fmt" |
| 7 | + "io" |
| 8 | + "net/http" |
| 9 | + "os" |
| 10 | + "path/filepath" |
| 11 | + "regexp" |
| 12 | + "sort" |
| 13 | + "strconv" |
| 14 | + "strings" |
| 15 | +) |
| 16 | + |
| 17 | +const ( |
| 18 | + textURL = "https://tanzil.net/pub/download/index.php?quranType=uthmani&outType=txt-2&agree=true&marks=true&sajdah=true&tatweel=true" |
| 19 | + metaURL = "https://tanzil.net/res/text/metadata/quran-data.js" |
| 20 | +) |
| 21 | + |
| 22 | +type ayah struct { |
| 23 | + Surah int `json:"surah"` |
| 24 | + SurahName string `json:"surah_name"` |
| 25 | + Ayah int `json:"ayah"` |
| 26 | + Juz int `json:"juz"` |
| 27 | + TextAR string `json:"text_ar"` |
| 28 | +} |
| 29 | + |
| 30 | +type point struct { |
| 31 | + Surah int |
| 32 | + Ayah int |
| 33 | +} |
| 34 | + |
| 35 | +func main() { |
| 36 | + if err := run(); err != nil { |
| 37 | + fmt.Fprintf(os.Stderr, "import failed: %v\n", err) |
| 38 | + os.Exit(1) |
| 39 | + } |
| 40 | +} |
| 41 | + |
| 42 | +func run() error { |
| 43 | + metadata, err := fetch(metaURL) |
| 44 | + if err != nil { |
| 45 | + return fmt.Errorf("fetch metadata: %w", err) |
| 46 | + } |
| 47 | + |
| 48 | + surahNames, err := parseSurahNames(metadata) |
| 49 | + if err != nil { |
| 50 | + return fmt.Errorf("parse surah names: %w", err) |
| 51 | + } |
| 52 | + |
| 53 | + juzStarts, err := parseJuzStarts(metadata) |
| 54 | + if err != nil { |
| 55 | + return fmt.Errorf("parse juz starts: %w", err) |
| 56 | + } |
| 57 | + |
| 58 | + textData, err := fetch(textURL) |
| 59 | + if err != nil { |
| 60 | + return fmt.Errorf("fetch quran text: %w", err) |
| 61 | + } |
| 62 | + |
| 63 | + ayahs, err := parseAyahLines(textData, surahNames, juzStarts) |
| 64 | + if err != nil { |
| 65 | + return fmt.Errorf("parse ayah lines: %w", err) |
| 66 | + } |
| 67 | + |
| 68 | + if err := validate(ayahs); err != nil { |
| 69 | + return err |
| 70 | + } |
| 71 | + |
| 72 | + outPath := filepath.Join("data", "quran.json") |
| 73 | + if err := writeJSON(outPath, ayahs); err != nil { |
| 74 | + return fmt.Errorf("write dataset: %w", err) |
| 75 | + } |
| 76 | + |
| 77 | + fmt.Printf("Imported %d ayahs into %s\n", len(ayahs), outPath) |
| 78 | + return nil |
| 79 | +} |
| 80 | + |
| 81 | +func fetch(url string) (string, error) { |
| 82 | + resp, err := http.Get(url) |
| 83 | + if err != nil { |
| 84 | + return "", err |
| 85 | + } |
| 86 | + defer resp.Body.Close() |
| 87 | + |
| 88 | + if resp.StatusCode != http.StatusOK { |
| 89 | + return "", fmt.Errorf("unexpected status: %s", resp.Status) |
| 90 | + } |
| 91 | + |
| 92 | + b, err := io.ReadAll(resp.Body) |
| 93 | + if err != nil { |
| 94 | + return "", err |
| 95 | + } |
| 96 | + |
| 97 | + return string(b), nil |
| 98 | +} |
| 99 | + |
| 100 | +func parseSurahNames(meta string) (map[int]string, error) { |
| 101 | + section, err := extractArray(meta, "QuranData.Sura") |
| 102 | + if err != nil { |
| 103 | + return nil, err |
| 104 | + } |
| 105 | + |
| 106 | + re := regexp.MustCompile(`(?m)^\s*\[(\d+),\s*\d+,\s*\d+,\s*\d+,\s*'[^']*',\s*"([^"]+)"`) |
| 107 | + matches := re.FindAllStringSubmatch(section, -1) |
| 108 | + if len(matches) < 114 { |
| 109 | + return nil, fmt.Errorf("expected 114 surah rows, got %d", len(matches)) |
| 110 | + } |
| 111 | + |
| 112 | + names := make(map[int]string, 114) |
| 113 | + for idx, m := range matches { |
| 114 | + n := idx + 1 |
| 115 | + names[n] = strings.TrimSpace(m[2]) |
| 116 | + } |
| 117 | + return names, nil |
| 118 | +} |
| 119 | + |
| 120 | +func parseJuzStarts(meta string) ([]point, error) { |
| 121 | + section, err := extractArray(meta, "QuranData.Juz") |
| 122 | + if err != nil { |
| 123 | + return nil, err |
| 124 | + } |
| 125 | + |
| 126 | + re := regexp.MustCompile(`\[(\d+),\s*(\d+)\]`) |
| 127 | + matches := re.FindAllStringSubmatch(section, -1) |
| 128 | + if len(matches) < 31 { |
| 129 | + return nil, fmt.Errorf("expected >=31 juz points, got %d", len(matches)) |
| 130 | + } |
| 131 | + |
| 132 | + out := make([]point, 0, len(matches)) |
| 133 | + for _, m := range matches { |
| 134 | + s, _ := strconv.Atoi(m[1]) |
| 135 | + a, _ := strconv.Atoi(m[2]) |
| 136 | + out = append(out, point{Surah: s, Ayah: a}) |
| 137 | + } |
| 138 | + |
| 139 | + return out, nil |
| 140 | +} |
| 141 | + |
| 142 | +func extractArray(meta, marker string) (string, error) { |
| 143 | + start := strings.Index(meta, marker) |
| 144 | + if start == -1 { |
| 145 | + return "", fmt.Errorf("marker not found: %s", marker) |
| 146 | + } |
| 147 | + from := strings.Index(meta[start:], "[") |
| 148 | + if from == -1 { |
| 149 | + return "", fmt.Errorf("array start not found for: %s", marker) |
| 150 | + } |
| 151 | + idx := start + from |
| 152 | + |
| 153 | + depth := 0 |
| 154 | + for i := idx; i < len(meta); i++ { |
| 155 | + switch meta[i] { |
| 156 | + case '[': |
| 157 | + depth++ |
| 158 | + case ']': |
| 159 | + depth-- |
| 160 | + if depth == 0 { |
| 161 | + return meta[idx : i+1], nil |
| 162 | + } |
| 163 | + } |
| 164 | + } |
| 165 | + |
| 166 | + return "", fmt.Errorf("array end not found for: %s", marker) |
| 167 | +} |
| 168 | + |
| 169 | +func parseAyahLines(text string, names map[int]string, juzStarts []point) ([]ayah, error) { |
| 170 | + scanner := bufio.NewScanner(strings.NewReader(text)) |
| 171 | + scanner.Buffer(make([]byte, 0, 64*1024), 2*1024*1024) |
| 172 | + |
| 173 | + out := make([]ayah, 0, 6236) |
| 174 | + for scanner.Scan() { |
| 175 | + line := strings.TrimSpace(scanner.Text()) |
| 176 | + if line == "" || strings.HasPrefix(line, "#") { |
| 177 | + continue |
| 178 | + } |
| 179 | + if !strings.Contains(line, "|") { |
| 180 | + continue |
| 181 | + } |
| 182 | + parts := strings.SplitN(line, "|", 3) |
| 183 | + if len(parts) != 3 { |
| 184 | + return nil, fmt.Errorf("invalid text line: %q", line) |
| 185 | + } |
| 186 | + |
| 187 | + surah, err := strconv.Atoi(parts[0]) |
| 188 | + if err != nil { |
| 189 | + return nil, fmt.Errorf("invalid surah in line %q", line) |
| 190 | + } |
| 191 | + ayahNum, err := strconv.Atoi(parts[1]) |
| 192 | + if err != nil { |
| 193 | + return nil, fmt.Errorf("invalid ayah in line %q", line) |
| 194 | + } |
| 195 | + |
| 196 | + name := names[surah] |
| 197 | + if name == "" { |
| 198 | + return nil, fmt.Errorf("missing surah name for surah %d", surah) |
| 199 | + } |
| 200 | + |
| 201 | + out = append(out, ayah{ |
| 202 | + Surah: surah, |
| 203 | + SurahName: name, |
| 204 | + Ayah: ayahNum, |
| 205 | + Juz: determineJuz(surah, ayahNum, juzStarts), |
| 206 | + TextAR: parts[2], |
| 207 | + }) |
| 208 | + } |
| 209 | + |
| 210 | + if err := scanner.Err(); err != nil { |
| 211 | + return nil, err |
| 212 | + } |
| 213 | + |
| 214 | + sort.Slice(out, func(i, j int) bool { |
| 215 | + if out[i].Surah == out[j].Surah { |
| 216 | + return out[i].Ayah < out[j].Ayah |
| 217 | + } |
| 218 | + return out[i].Surah < out[j].Surah |
| 219 | + }) |
| 220 | + |
| 221 | + return out, nil |
| 222 | +} |
| 223 | + |
| 224 | +func determineJuz(surah, ayah int, starts []point) int { |
| 225 | + for i := 0; i < len(starts)-1; i++ { |
| 226 | + if before(surah, ayah, starts[i+1].Surah, starts[i+1].Ayah) { |
| 227 | + return i + 1 |
| 228 | + } |
| 229 | + } |
| 230 | + return 30 |
| 231 | +} |
| 232 | + |
| 233 | +func before(s1, a1, s2, a2 int) bool { |
| 234 | + if s1 != s2 { |
| 235 | + return s1 < s2 |
| 236 | + } |
| 237 | + return a1 < a2 |
| 238 | +} |
| 239 | + |
| 240 | +func validate(ayahs []ayah) error { |
| 241 | + if len(ayahs) != 6236 { |
| 242 | + return fmt.Errorf("expected 6236 ayahs, got %d", len(ayahs)) |
| 243 | + } |
| 244 | + |
| 245 | + seen := make(map[string]struct{}, len(ayahs)) |
| 246 | + for _, a := range ayahs { |
| 247 | + if a.Surah < 1 || a.Surah > 114 { |
| 248 | + return fmt.Errorf("invalid surah: %d", a.Surah) |
| 249 | + } |
| 250 | + if a.Ayah < 1 { |
| 251 | + return fmt.Errorf("invalid ayah: %d", a.Ayah) |
| 252 | + } |
| 253 | + if a.Juz < 1 || a.Juz > 30 { |
| 254 | + return fmt.Errorf("invalid juz: %d for %d:%d", a.Juz, a.Surah, a.Ayah) |
| 255 | + } |
| 256 | + if strings.TrimSpace(a.TextAR) == "" { |
| 257 | + return fmt.Errorf("empty text_ar for %d:%d", a.Surah, a.Ayah) |
| 258 | + } |
| 259 | + |
| 260 | + k := fmt.Sprintf("%d:%d", a.Surah, a.Ayah) |
| 261 | + if _, ok := seen[k]; ok { |
| 262 | + return fmt.Errorf("duplicate ayah key: %s", k) |
| 263 | + } |
| 264 | + seen[k] = struct{}{} |
| 265 | + } |
| 266 | + |
| 267 | + return nil |
| 268 | +} |
| 269 | + |
| 270 | +func writeJSON(path string, v any) error { |
| 271 | + if err := os.MkdirAll(filepath.Dir(path), 0o755); err != nil { |
| 272 | + return err |
| 273 | + } |
| 274 | + |
| 275 | + f, err := os.Create(path) |
| 276 | + if err != nil { |
| 277 | + return err |
| 278 | + } |
| 279 | + defer f.Close() |
| 280 | + |
| 281 | + enc := json.NewEncoder(f) |
| 282 | + enc.SetIndent("", " ") |
| 283 | + enc.SetEscapeHTML(false) |
| 284 | + return enc.Encode(v) |
| 285 | +} |
0 commit comments