| 1 | package AmberDB::Base::Schema; | ||
| 2 | |||
| 3 | use 5.016; | ||
| 4 | use warnings; | ||
| 5 | use Carp qw(croak cluck); | ||
| 6 | use File::Spec; | ||
| 7 | use parent 'AmberDB::Base'; | ||
| 8 | |||
| 9 | our $VERSION = '5.26.6'; | ||
| 10 | my $CREATED = '2026-09-06'; | ||
| 11 | |||
| 12 | # Whitelist of inheritable group-level configuration keys from .dbase to member tables | ||
| 13 | our %INHERITABLE_DBASE_KEYS = map { $_ => 1 } qw( | ||
| 14 | use_ramdisk ramdisk_ttl use_cache cache_ttl | ||
| 15 | keep_deleted log_owner use_counter | ||
| 16 | use_section section use_year year | ||
| 17 | use_language lang | ||
| 18 | no_transact no_backup table_dir use_junk | ||
| 19 | use_simple | ||
| 20 | ); | ||
| 21 | |||
| 22 | # ===================================================================== | ||
| 23 | # SCHEMA FIELD NORMALIZATION, VALIDATION & TYPE CONVERSION | ||
| 24 | # ===================================================================== | ||
| 25 | |||
| 26 | # Normalizes and validates a single field value according to schema block definition. | ||
| 27 | # Usage: | ||
| 28 | # my $clean_val = $adb->enc_field($blk_def, $val); | ||
| 29 | # ------------------------------------------------ | ||
| 30 | sub enc_field { | ||
| 31 | my ( $self, $blk, $val ) = @_; | ||
| 32 | |||
| 33 | if ( $blk && ref($blk) eq 'HASH' ) { | ||
| 34 | my $type = lc( $blk->{type} // 'text' ); | ||
| 35 | my $valid = lc( $blk->{valid} // '' ); | ||
| 36 | |||
| 37 | # 1. Type: number / num (integer, float, negative support) | ||
| 38 | if ( $type eq 'num' || $type eq 'number' || $type eq 'numeric' || $type eq 'int' || $type eq 'float' || $type eq 'decimal' ) { | ||
| 39 | if ( defined $val && length("$val") ) { | ||
| 40 | ( my $trimmed = "$val" ) =~ s/^\s+|\s+$//g; | ||
| 41 | if ( $trimmed =~ /^[+-]?[0-9]+(?:\.[0-9]+)?$/ ) { | ||
| 42 | $val = 0 + $trimmed; | ||
| 43 | } | ||
| 44 | else { | ||
| 45 | $val = 0; | ||
| 46 | } | ||
| 47 | } | ||
| 48 | else { | ||
| 49 | $val = 0; | ||
| 50 | } | ||
| 51 | } | ||
| 52 | # 2. Type: ascii | ||
| 53 | elsif ( $type eq 'ascii' ) { | ||
| 54 | if ( defined $val && length("$val") ) { | ||
| 55 | $val = $self->to_ascii("$val"); | ||
| 56 | } | ||
| 57 | else { | ||
| 58 | $val = ''; | ||
| 59 | } | ||
| 60 | } | ||
| 61 | # 3. Type: date | ||
| 62 | elsif ( $type eq 'date' || $type eq 'datetime' || $type eq 'date_short' || $type eq 'date_long' ) { | ||
| 63 | if ( ( !defined $val || $val eq '' ) && $valid =~ /auto_date/ ) { | ||
| 64 | my $y = $self->year; | ||
| 65 | my $m = $self->month; | ||
| 66 | my $d = $self->day; | ||
| 67 | $val = "$y-$m-$d"; | ||
| 68 | } | ||
| 69 | else { | ||
| 70 | $val //= ''; | ||
| 71 | } | ||
| 72 | } | ||
| 73 | # 4. Type: array / repeat / loop | ||
| 74 | elsif ( $type eq 'array' || $type eq 'list' || $type eq 'repeat' || $type eq 'repeats' || $type eq 'loop' ) { | ||
| 75 | if ( defined $val ) { | ||
| 76 | if ( ref($val) eq 'ARRAY' ) { | ||
| 77 | # ok | ||
| 78 | } | ||
| 79 | elsif ( ref($val) ) { | ||
| 80 | $val = [$val]; | ||
| 81 | } | ||
| 82 | elsif ( length("$val") ) { | ||
| 83 | $val = [ split /[,|]/, "$val" ]; | ||
| 84 | } | ||
| 85 | else { | ||
| 86 | $val = []; | ||
| 87 | } | ||
| 88 | } | ||
| 89 | else { | ||
| 90 | $val = []; | ||
| 91 | } | ||
| 92 | } | ||
| 93 | # 5. Type: hash | ||
| 94 | elsif ( $type eq 'hash' || $type eq 'dict' || $type eq 'json' ) { | ||
| 95 | if ( defined $val && ref($val) eq 'HASH' ) { | ||
| 96 | # ok | ||
| 97 | } | ||
| 98 | else { | ||
| 99 | $val = {}; | ||
| 100 | } | ||
| 101 | } | ||
| 102 | # 6. Type: text / string | ||
| 103 | elsif ( $type eq 'text' || $type eq 'string' || $type eq 'tinytext' || $type eq 'html' ) { | ||
| 104 | $val //= ''; | ||
| 105 | } | ||
| 106 | # 7. Type: binary / base64 | ||
| 107 | elsif ( $type eq 'binary' || $type eq 'base64' ) { | ||
| 108 | $val //= ''; | ||
| 109 | } | ||
| 110 | # 8. Type: auto_id | ||
| 111 | elsif ( $type eq 'auto_id' || $type eq 'autoid' ) { | ||
| 112 | # auto_id is handled by table_autoid | ||
| 113 | } | ||
| 114 | } | ||
| 115 | |||
| 116 | return $val; | ||
| 117 | } | ||
| 118 | |||
| 119 | # Normalizes and validates field values according to schema block definitions prior to encoding/writing. | ||
| 120 | # Usage: | ||
| 121 | # my @clean_fields = $adb->enc_validate($tableid, \@fields, $has_id); | ||
| 122 | # ------------------------------------------------ | ||
| 123 | sub enc_validate { | ||
| 124 | my ( $self, $tableid, $fields_ref, $has_id ) = @_; | ||
| 125 | |||
| 126 | return wantarray ? () : [] unless defined $fields_ref; | ||
| 127 | my @fields = ( ref($fields_ref) eq 'ARRAY' ) ? @$fields_ref : ($fields_ref); | ||
| 128 | return wantarray ? @fields : \@fields unless @fields; | ||
| 129 | |||
| 130 | # Bypass in simple mode or when no tableid is given | ||
| 131 | return wantarray ? @fields : \@fields | ||
| 132 | if $self->config('use_simple') || !defined $tableid || $tableid eq ''; | ||
| 133 | |||
| 134 | my $table_info = $self->table_info($tableid); | ||
| 135 | return wantarray ? @fields : \@fields | ||
| 136 | unless $table_info && ref( $table_info->{blocks} ) eq 'ARRAY' && @{ $table_info->{blocks} }; | ||
| 137 | |||
| 138 | my $blocks = $table_info->{blocks}; | ||
| 139 | my @cleaned; | ||
| 140 | |||
| 141 | for ( my $i = 0 ; $i < @fields ; $i++ ) { | ||
| 142 | my $blk_idx = $has_id ? $i : ( $i + 1 ); | ||
| 143 | my $blk = $blocks->[$blk_idx]; | ||
| 144 | if ( !$blk && $table_info->{repeat_start} && $blk_idx >= $table_info->{repeat_start} ) { | ||
| 145 | $blk = $blocks->[ $table_info->{repeat_start} ] // $blocks->[-1]; | ||
| 146 | } | ||
| 147 | push @cleaned, $self->enc_field( $blk, $fields[$i] ); | ||
| 148 | } | ||
| 149 | |||
| 150 | return wantarray ? @cleaned : \@cleaned; | ||
| 151 | } | ||
| 152 | |||
| 153 | # Normalizes and casts a single field value according to schema block definition upon decoding/reading. | ||
| 154 | # Usage: | ||
| 155 | # my $decoded_val = $adb->dec_field($blk_def, $val); | ||
| 156 | # ------------------------------------------------ | ||
| 157 | sub dec_field { | ||
| 158 | my ( $self, $blk, $val ) = @_; | ||
| 159 | |||
| 160 | if ( $blk && ref($blk) eq 'HASH' ) { | ||
| 161 | my $type = lc( $blk->{type} // 'text' ); | ||
| 162 | |||
| 163 | # 1. Type: number / num -> Ensure numeric scalar (0 + $val) | ||
| 164 | if ( $type eq 'num' || $type eq 'number' || $type eq 'numeric' || $type eq 'int' || $type eq 'float' || $type eq 'decimal' ) { | ||
| 165 | if ( defined $val && length("$val") ) { | ||
| 166 | ( my $trimmed = "$val" ) =~ s/^\s+|\s+$//g; | ||
| 167 | if ( $trimmed =~ /^[+-]?[0-9]+(?:\.[0-9]+)?$/ ) { | ||
| 168 | $val = 0 + $trimmed; | ||
| 169 | } | ||
| 170 | else { | ||
| 171 | $val = 0; | ||
| 172 | } | ||
| 173 | } | ||
| 174 | else { | ||
| 175 | $val = 0; | ||
| 176 | } | ||
| 177 | } | ||
| 178 | # 2. Type: array / repeat / loop -> Ensure ARRAY ref | ||
| 179 | elsif ( $type eq 'array' || $type eq 'list' || $type eq 'repeat' || $type eq 'repeats' || $type eq 'loop' ) { | ||
| 180 | if ( !defined $val ) { | ||
| 181 | $val = []; | ||
| 182 | } | ||
| 183 | elsif ( ref($val) ne 'ARRAY' ) { | ||
| 184 | $val = length("$val") ? [ split /[,|]/, "$val" ] : []; | ||
| 185 | } | ||
| 186 | } | ||
| 187 | # 3. Type: hash -> Ensure HASH ref | ||
| 188 | elsif ( $type eq 'hash' || $type eq 'dict' || $type eq 'json' ) { | ||
| 189 | if ( !defined $val || ref($val) ne 'HASH' ) { | ||
| 190 | $val = {}; | ||
| 191 | } | ||
| 192 | } | ||
| 193 | # 4. Text / ASCII / Date / Binary | ||
| 194 | else { | ||
| 195 | $val //= ''; | ||
| 196 | } | ||
| 197 | } | ||
| 198 | |||
| 199 | return $val; | ||
| 200 | } | ||
| 201 | |||
| 202 | # Normalizes and casts field values according to schema block definitions upon decoding/reading. | ||
| 203 | # Usage: | ||
| 204 | # my @decoded_fields = $adb->dec_validate($tableid, \@fields, $has_id); | ||
| 205 | # ------------------------------------------------ | ||
| 206 | sub dec_validate { | ||
| 207 | my ( $self, $tableid, $fields_ref, $has_id ) = @_; | ||
| 208 | |||
| 209 | return wantarray ? () : [] unless defined $fields_ref; | ||
| 210 | my @fields = ( ref($fields_ref) eq 'ARRAY' ) ? @$fields_ref : ($fields_ref); | ||
| 211 | return wantarray ? @fields : \@fields unless @fields; | ||
| 212 | |||
| 213 | # Bypass in simple mode or when no tableid is given | ||
| 214 | return wantarray ? @fields : \@fields | ||
| 215 | if $self->config('use_simple') || !defined $tableid || $tableid eq ''; | ||
| 216 | |||
| 217 | my $table_info = $self->table_info($tableid); | ||
| 218 | return wantarray ? @fields : \@fields | ||
| 219 | unless $table_info && ref( $table_info->{blocks} ) eq 'ARRAY' && @{ $table_info->{blocks} }; | ||
| 220 | |||
| 221 | my $blocks = $table_info->{blocks}; | ||
| 222 | my @cleaned; | ||
| 223 | |||
| 224 | for ( my $i = 0 ; $i < @fields ; $i++ ) { | ||
| 225 | my $blk_idx = $has_id ? $i : ( $i + 1 ); | ||
| 226 | my $blk = $blocks->[$blk_idx]; | ||
| 227 | if ( !$blk && $table_info->{repeat_start} && $blk_idx >= $table_info->{repeat_start} ) { | ||
| 228 | $blk = $blocks->[ $table_info->{repeat_start} ] // $blocks->[-1]; | ||
| 229 | } | ||
| 230 | push @cleaned, $self->dec_field( $blk, $fields[$i] ); | ||
| 231 | } | ||
| 232 | |||
| 233 | return wantarray ? @cleaned : \@cleaned; | ||
| 234 | } | ||
| 235 | |||
| 236 | # ===================================================================== | ||
| 237 | # TABLE SANITIZATION & SCHEMA ARGUMENTS | ||
| 238 | # ===================================================================== | ||
| 239 | |||
| 240 | # Sanitizes table/file identifier. | ||
| 241 | # Allows subdirectories and dots (e.g. 'uyeler/bekleyen.uyeler', '/.dosya') | ||
| 242 | # Strictly strips path traversal segments ('..', '.', './', '.\', '../', '..\') | ||
| 243 | # ------------------------------------------------ | ||
| 244 | sub sanitize_table { | ||
| 245 | my ( $self, $table ) = @_; | ||
| 246 | return "" unless defined $table && length $table; | ||
| 247 | |||
| 248 | # 1. Normalize backslashes to forward slashes | ||
| 249 | $table =~ s/\\/\//g; | ||
| 250 | |||
| 251 | # 2. Strip database extension if present at end | ||
| 252 | my $ext = $self->{db_ext} // "db"; | ||
| 253 | $table =~ s/\.\Q$ext\E$//; | ||
| 254 | |||
| 255 | # 3. Filter allowed characters: alphanumeric, underscore, dot, hyphen, slash | ||
| 256 | $table =~ s/[^\w\.\-\/]+//g; | ||
| 257 | |||
| 258 | # 4. Collapse multiple consecutive slashes | ||
| 259 | $table =~ s{/+}{/}g; | ||
| 260 | |||
| 261 | # 5. Remove path traversal components: split into segments and filter out '.' and '..' | ||
| 262 | my @clean_segments; | ||
| 263 | foreach my $seg ( split m{/}, $table ) { | ||
| 264 | next if $seg eq '' || $seg eq '.' || $seg eq '..'; | ||
| 265 | push @clean_segments, $seg; | ||
| 266 | } | ||
| 267 | |||
| 268 | return join( '/', @clean_segments ); | ||
| 269 | } | ||
| 270 | |||
| 271 | # ($id, $path) = $self->schema_arg($arg, "table"|"dbase") | ||
| 272 | # ------------------------------------------------ | ||
| 273 | sub schema_arg { | ||
| 274 | my ( $self, $arg, $ext ) = @_; | ||
| 275 | return ( "", "" ) unless defined $arg && length $arg; | ||
| 276 | |||
| 277 | if ( $arg =~ m{[\\/]|\.${ext}$} ) { | ||
| 278 | ( my $id = $arg ) =~ s{.+[\\/]([^\\/]+?)(?:\.${ext})?$}{$1}; | ||
| 279 | $id = $self->sanitize_table($id); | ||
| 280 | return ( $id, $arg ); | ||
| 281 | } | ||
| 282 | my $clean = $self->sanitize_table($arg); | ||
| 283 | my $schema_dir = $self->path('schema_dir') | ||
| 284 | || ( $self->path('dbase_dir') ? $self->path('dbase_dir') . "/schema" : "schema" ); | ||
| 285 | return ( $clean, "$schema_dir/$clean.$ext" ); | ||
| 286 | } | ||
| 287 | |||
| 288 | # ===================================================================== | ||
| 289 | # SCHEMA LOADERS, ATTRIBUTES & WRITERS | ||
| 290 | # ===================================================================== | ||
| 291 | |||
| 292 | # Safely evaluates a Perl schema definition file via `do`. | ||
| 293 | # Normalizes paths, catches syntax errors and verifies hash structure. | ||
| 294 | # --------------------------------------------------------------------------- | ||
| 295 | sub do_schema { | ||
| 296 | my ( $self, $path ) = @_; | ||
| 297 | return unless defined $path && length $path && -f $path; | ||
| 298 | |||
| 299 | $path =~ s{\\}{/}g; | ||
| 300 | $path = "./$path" unless $path =~ m{^(?:\./|/|[a-zA-Z]:)}; | ||
| 301 | |||
| 302 | my $data = eval { do $path }; | ||
| 303 | if ($@) { | ||
| 304 | cluck "[AMBERDB_SCHEMA] Syntax error in schema file '$path': $@\n"; | ||
| 305 | return undef; | ||
| 306 | } | ||
| 307 | |||
| 308 | return ( ref($data) eq 'HASH' ) ? $data : undef; | ||
| 309 | } | ||
| 310 | |||
| 311 | # Retrieves database group schema definition. | ||
| 312 | # my $dbase_info = $adb->dbase_info($dbase); | ||
| 313 | # ------------------------------------------------ | ||
| 314 | sub dbase_info { | ||
| 315 | my ( $self, $arg ) = @_; | ||
| 316 | |||
| 317 | return unless $arg; | ||
| 318 | my ( $dbase, $dbase_path ) = $self->schema_arg( $arg, "dbase" ); | ||
| 319 | |||
| 320 | return $self->{_dbase}->{$dbase} | ||
| 321 | if $self->{_dbase}->{$dbase} && %{ $self->{_dbase}->{$dbase} }; | ||
| 322 | |||
| 323 | my $ramdisk_schema = $self->{_path}->{schema_rdir} || $self->path('schema_rdir'); | ||
| 324 | my $target_path = $dbase_path; | ||
| 325 | if ( $ramdisk_schema && -e "$ramdisk_schema/$dbase.dbase" ) { | ||
| 326 | $target_path = "$ramdisk_schema/$dbase.dbase"; | ||
| 327 | } | ||
| 328 | |||
| 329 | if ( my $do_data = $self->do_schema($target_path) ) { | ||
| 330 | $do_data->{name} = $self->utf_decode( $do_data->{name} ) if defined $do_data->{name}; | ||
| 331 | $self->{_dbase}->{$dbase} = $do_data; | ||
| 332 | if ( $ramdisk_schema && !-e "$ramdisk_schema/$dbase.dbase" && $self->dir_exist($ramdisk_schema) ) { | ||
| 333 | require File::Copy; | ||
| 334 | eval { File::Copy::copy( $dbase_path, "$ramdisk_schema/$dbase.dbase" ) }; | ||
| 335 | } | ||
| 336 | } | ||
| 337 | |||
| 338 | return $self->{_dbase}->{$dbase}; | ||
| 339 | } | ||
| 340 | |||
| 341 | # Applies inherited configuration from the database group (.dbase) to a table schema. | ||
| 342 | # Only copies whitelisted keys (%INHERITABLE_DBASE_KEYS) that are not already defined | ||
| 343 | # on the table schema (preserving table-level overrides). | ||
| 344 | sub _apply_dbase_inheritance { | ||
| 345 | my ( $self, $table, $schema ) = @_; | ||
| 346 | |||
| 347 | return unless ref($schema) eq 'HASH'; | ||
| 348 | |||
| 349 | my $dbase = $schema->{dbase}; | ||
| 350 | if ( !defined $dbase || !length $dbase ) { | ||
| 351 | my $t = defined $table ? $table : ( $schema->{table} // '' ); | ||
| 352 | ($dbase) = ( $t =~ /^([a-z0-9]+)_/i ); | ||
| 353 | $dbase //= ""; | ||
| 354 | } | ||
| 355 | return unless length $dbase; | ||
| 356 | |||
| 357 | my $dbase_info = $self->dbase_info($dbase); | ||
| 358 | return unless $dbase_info && ref($dbase_info) eq 'HASH'; | ||
| 359 | |||
| 360 | for my $key ( keys %INHERITABLE_DBASE_KEYS ) { | ||
| 361 | if ( exists $dbase_info->{$key} && !exists $schema->{$key} ) { | ||
| 362 | $schema->{$key} = $dbase_info->{$key}; | ||
| 363 | } | ||
| 364 | } | ||
| 365 | } | ||
| 366 | |||
| 367 | # ============================================================================ | ||
| 368 | # TABLE SCHEMA NORMALIZATION & VALIDATION PIPELINE | ||
| 369 | # Normalizes, converts, and validates table schema definitions through an ordered | ||
| 370 | # rule pipeline. Ensures consistent in-memory structures across disk loading, | ||
| 371 | # table_attr, and table_infset. | ||
| 372 | # Usage: | ||
| 373 | # $adb->normalize_blocks($table, $schema_hashref); | ||
| 374 | # ============================================================================ | ||
| 375 | sub normalize_blocks { | ||
| 376 | my ( $self, $table, $schema ) = @_; | ||
| 377 | |||
| 378 | return {} unless ref($schema) eq 'HASH'; | ||
| 379 | |||
| 380 | # ------------------------------------------------------------------------ | ||
| 381 | # Pipeline Step 0: Table Identity & Metadata | ||
| 382 | # Ensure table name and database group are present in schema. | ||
| 383 | # ------------------------------------------------------------------------ | ||
| 384 | if ( defined $table && length $table ) { | ||
| 385 | $schema->{table} = $table unless defined $schema->{table} && length $schema->{table}; | ||
| 386 | my ($dbase) = ( $table =~ /^([a-z0-9]+)_/i ); | ||
| 387 | $dbase //= ""; | ||
| 388 | $schema->{dbase} = $dbase unless defined $schema->{dbase} && length $schema->{dbase}; | ||
| 389 | } | ||
| 390 | |||
| 391 | # Normalize UTF-8 characters on schema identity and block definitions | ||
| 392 | $schema->{name} = $self->utf_decode( $schema->{name} ) if defined $schema->{name}; | ||
| 393 | if ( ref( $schema->{blocks} ) eq 'ARRAY' ) { | ||
| 394 | for my $b ( @{ $schema->{blocks} } ) { | ||
| 395 | next unless ref($b) eq 'HASH'; | ||
| 396 | for my $k (qw(name option valid)) { | ||
| 397 | $b->{$k} = $self->utf_decode( $b->{$k} ) if defined $b->{$k}; | ||
| 398 | } | ||
| 399 | } | ||
| 400 | } | ||
| 401 | if ( ref( $schema->{facet_block} ) eq 'ARRAY' ) { | ||
| 402 | for my $fb ( @{ $schema->{facet_block} } ) { | ||
| 403 | next unless ref($fb) eq 'HASH'; | ||
| 404 | for my $k (qw(label name title)) { | ||
| 405 | $fb->{$k} = $self->utf_decode( $fb->{$k} ) if defined $fb->{$k}; | ||
| 406 | } | ||
| 407 | } | ||
| 408 | } | ||
| 409 | |||
| 410 | # ------------------------------------------------------------------------ | ||
| 411 | # Pipeline Step 0b: Dbase Schema Group Inheritance | ||
| 412 | # Inherit whitelisted operational flags from .dbase if not overridden. | ||
| 413 | # ------------------------------------------------------------------------ | ||
| 414 | $self->_apply_dbase_inheritance( $table, $schema ); | ||
| 415 | |||
| 416 | # ------------------------------------------------------------------------ | ||
| 417 | # Pipeline Step 1: Global Configuration Inheritance | ||
| 418 | # ------------------------------------------------------------------------ | ||
| 419 | if ( defined $schema->{use_ramdisk} ) { | ||
| 420 | $schema->{use_ramdisk} = $self->_normalize_ramdisk_tier( $schema->{use_ramdisk} ); | ||
| 421 | } | ||
| 422 | elsif ( !defined $schema->{use_cache} ) { | ||
| 423 | my $global_ram = $self->config('use_ramdisk'); | ||
| 424 | $global_ram = $self->_normalize_ramdisk_tier($global_ram); | ||
| 425 | $global_ram = 0 if $global_ram == 3; | ||
| 426 | $schema->{use_ramdisk} = $global_ram; | ||
| 427 | } | ||
| 428 | |||
| 429 | # ------------------------------------------------------------------------ | ||
| 430 | # Pipeline Step 1b: Volatile RAM-Disk Tier 3 Stripping | ||
| 431 | # If use_ramdisk == 3, table is pure volatile key-value store on RAM-disk. | ||
| 432 | # Strip all indexing, columnar, facet, and relational definitions. | ||
| 433 | # ------------------------------------------------------------------------ | ||
| 434 | if ( ( $schema->{use_ramdisk} // 0 ) == 3 ) { | ||
| 435 | delete @{$schema}{ | ||
| 436 | qw( | ||
| 437 | blocks match_block search_block view_block facet_block filter_block | ||
| 438 | suggest_block suggest_join | ||
| 439 | slug_block sort_block sort_fields use_facet facet_rules use_junk junk_rules | ||
| 440 | record_index repeat_start repeat_ids field_rules keep_deleted | ||
| 441 | ) | ||
| 442 | }; | ||
| 443 | $schema->{use_simple} = 1; | ||
| 444 | $schema->{no_backup} = 1; | ||
| 445 | $schema->{no_transact} = 1; | ||
| 446 | $schema->{ramdisk_ttl} = 300 unless defined $schema->{ramdisk_ttl}; | ||
| 447 | } | ||
| 448 | |||
| 449 | # ------------------------------------------------------------------------ | ||
| 450 | # Pipeline Step 2: Simple Mode Stripping | ||
| 451 | # In simple key-value mode, selectively strip columnar, indexing, relational, | ||
| 452 | # and caching definitions to enforce lightweight operation. | ||
| 453 | # ------------------------------------------------------------------------ | ||
| 454 | if ( $schema->{use_simple} && ( $schema->{use_ramdisk} // 0 ) != 3 ) { | ||
| 455 | delete @{$schema}{ | ||
| 456 | qw( | ||
| 457 | blocks match_block search_block view_block facet_block filter_block | ||
| 458 | suggest_block suggest_join | ||
| 459 | slug_block sort_block sort_fields use_facet facet_rules use_junk junk_rules | ||
| 460 | record_index repeat_start repeat_ids field_rules | ||
| 461 | use_ramdisk ramdisk_ttl use_cache cache_ttl | ||
| 462 | ) | ||
| 463 | }; | ||
| 464 | return $schema; | ||
| 465 | } | ||
| 466 | |||
| 467 | # ------------------------------------------------------------------------ | ||
| 468 | # Pipeline Step 3: RAM-Disk Environment & Mount Verification | ||
| 469 | # If RAM-disk is not physically mounted or ready, strip use_ramdisk. | ||
| 470 | # The existence of $schema->{use_ramdisk} is a guaranteed contract of readiness. | ||
| 471 | # ------------------------------------------------------------------------ | ||
| 472 | unless ( $self->ramdisk_is_mounted() ) { | ||
| 473 | delete $schema->{use_ramdisk}; | ||
| 474 | delete $schema->{use_cache}; | ||
| 475 | } | ||
| 476 | |||
| 477 | # ------------------------------------------------------------------------ | ||
| 478 | # Pipeline Step 4: Blocks & Relational Foreign Keys (RDBM) Normalization | ||
| 479 | # Converts string RDBM definitions like: | ||
| 480 | # "catalog_category,2", "catalog_category;2", "catalog_category|2", "catalog_category:2" | ||
| 481 | # into canonical hashref: { table => "catalog_category", display => 2 } | ||
| 482 | # ------------------------------------------------------------------------ | ||
| 483 | if ( exists $schema->{blocks} ) { | ||
| 484 | if ( ref( $schema->{blocks} ) eq 'ARRAY' ) { | ||
| 485 | foreach my $block ( @{ $schema->{blocks} } ) { | ||
| 486 | $self->normalize_rdbm($block) if ref($block) eq 'HASH'; | ||
| 487 | } | ||
| 488 | } | ||
| 489 | elsif ( ref( $schema->{blocks} ) eq 'HASH' ) { | ||
| 490 | foreach my $key ( keys %{ $schema->{blocks} } ) { | ||
| 491 | my $block = $schema->{blocks}->{$key}; | ||
| 492 | $self->normalize_rdbm($block) if ref($block) eq 'HASH'; | ||
| 493 | } | ||
| 494 | } | ||
| 495 | } | ||
| 496 | |||
| 497 | # ------------------------------------------------------------------------ | ||
| 498 | # Pipeline Step 4b: Precompute _has_schema and _block_keys for high-performance lookup | ||
| 499 | # ------------------------------------------------------------------------ | ||
| 500 | if ( exists $schema->{blocks} && ref( $schema->{blocks} ) eq 'ARRAY' && @{ $schema->{blocks} } ) { | ||
| 501 | $schema->{_has_schema} = 1; | ||
| 502 | my %b_keys; | ||
| 503 | for my $idx ( 0 .. $#{ $schema->{blocks} } ) { | ||
| 504 | my $b = $schema->{blocks}->[$idx]; | ||
| 505 | $b_keys{$idx} = 1; | ||
| 506 | if ( ref($b) eq 'HASH' ) { | ||
| 507 | my $name = $b->{name}; | ||
| 508 | my $id = $b->{id}; | ||
| 509 | $b_keys{ lc($name) } = 1 if defined $name; | ||
| 510 | $b_keys{ lc($id) } = 1 if defined $id; | ||
| 511 | |||
| 512 | # Plural and singular aliases for repeating blocks | ||
| 513 | if ( defined $name ) { | ||
| 514 | $b_keys{ lc("${name}s") } = 1; | ||
| 515 | $b_keys{ lc($1) } = 1 if $name =~ /^(.+)s$/i; | ||
| 516 | } | ||
| 517 | if ( defined $id && ( !defined $name || $id ne $name ) ) { | ||
| 518 | $b_keys{ lc("${id}s") } = 1; | ||
| 519 | $b_keys{ lc($1) } = 1 if $id =~ /^(.+)s$/i; | ||
| 520 | } | ||
| 521 | } | ||
| 522 | } | ||
| 523 | $b_keys{repeat} = 1 if $schema->{repeat_start}; | ||
| 524 | $b_keys{repeats} = 1 if $schema->{repeat_start}; | ||
| 525 | $schema->{_block_keys} = \%b_keys; | ||
| 526 | } | ||
| 527 | else { | ||
| 528 | $schema->{_has_schema} = 0; | ||
| 529 | $schema->{_block_keys} = {}; | ||
| 530 | } | ||
| 531 | |||
| 532 | # ------------------------------------------------------------------------ | ||
| 533 | # Pipeline Step 5: (Future Schema Rules / Extensions) | ||
| 534 | # Alanları ve sınırları belirlenmiş yeni şema kuralları buraya eklenebilir. | ||
| 535 | # ------------------------------------------------------------------------ | ||
| 536 | |||
| 537 | return $schema; | ||
| 538 | } | ||
| 539 | |||
| 540 | # Normalizes rdbm definition within a block into canonical { table => $t, display => $d } | ||
| 541 | sub normalize_rdbm { | ||
| 542 | my ( $self, $block ) = @_; | ||
| 543 | return unless ref($block) eq 'HASH'; | ||
| 544 | |||
| 545 | return unless defined $block->{rdbm} && $block->{rdbm} ne ''; | ||
| 546 | |||
| 547 | # Case A: Already a hashref { table => '...', display => ... } | ||
| 548 | if ( ref( $block->{rdbm} ) eq 'HASH' ) { | ||
| 549 | my $t = $block->{rdbm}->{table}; | ||
| 550 | my $d = $block->{rdbm}->{display}; | ||
| 551 | $t = defined $t ? "$t" : ""; | ||
| 552 | $t =~ s/^\s+|\s+$//g; | ||
| 553 | $d = ( defined $d && $d =~ /^\d+$/ ) ? int($d) : ( defined $block->{display} && $block->{display} =~ /^\d+$/ ? int($block->{display}) : 1 ); | ||
| 554 | $block->{rdbm} = { | ||
| 555 | table => $t, | ||
| 556 | display => $d, | ||
| 557 | }; | ||
| 558 | return; | ||
| 559 | } | ||
| 560 | |||
| 561 | # Case B: String format like "catalog_category,2", "catalog_category;2", "catalog_category|2", "catalog_category:2" | ||
| 562 | if ( !ref( $block->{rdbm} ) ) { | ||
| 563 | my $raw = "$block->{rdbm}"; | ||
| 564 | $raw =~ s/^\s+|\s+$//g; | ||
| 565 | if ( !length $raw ) { | ||
| 566 | delete $block->{rdbm}; | ||
| 567 | return; | ||
| 568 | } | ||
| 569 | |||
| 570 | # Split on comma, semicolon, pipe, or colon (, ; | :) | ||
| 571 | my ( $t, $d ) = split( /[,;:|]/, $raw, 2 ); | ||
| 572 | $t =~ s/^\s+|\s+$//g if defined $t; | ||
| 573 | $d =~ s/^\s+|\s+$//g if defined $d; | ||
| 574 | |||
| 575 | if ( defined $t && length $t ) { | ||
| 576 | my $disp = ( defined $d && $d =~ /^\d+$/ ) | ||
| 577 | ? int($d) | ||
| 578 | : ( defined $block->{display} && $block->{display} =~ /^\d+$/ | ||
| 579 | ? int( $block->{display} ) | ||
| 580 | : 1 ); | ||
| 581 | $block->{rdbm} = { | ||
| 582 | table => $t, | ||
| 583 | display => $disp, | ||
| 584 | }; | ||
| 585 | } | ||
| 586 | else { | ||
| 587 | delete $block->{rdbm}; | ||
| 588 | } | ||
| 589 | } | ||
| 590 | } | ||
| 591 | |||
| 592 | # my $table_info = $adb->table_info($table_id); | ||
| 593 | # ------------------------------------------------ | ||
| 594 | sub table_info { | ||
| 595 | my ( $self, $arg ) = @_; | ||
| 596 | |||
| 597 | return {} unless $arg; | ||
| 598 | return {} if $self->config('use_simple'); | ||
| 599 | |||
| 600 | my ( $table, $table_path ) = $self->schema_arg( $arg, "table" ); | ||
| 601 | $table && $table_path or return {}; | ||
| 602 | |||
| 603 | if ( $self->{_table}->{$table} && %{ $self->{_table}->{$table} } ) { | ||
| 604 | return { %{ $self->{_table}->{$table} } }; | ||
| 605 | } | ||
| 606 | |||
| 607 | my ($dbase) = ( $table =~ /^([a-z0-9]+)_/i ); | ||
| 608 | $dbase //= ""; | ||
| 609 | |||
| 610 | my $ramdisk_schema = $self->path('schema_rdir'); | ||
| 611 | my $target_path = $table_path; | ||
| 612 | if ( $ramdisk_schema && -e "$ramdisk_schema/$table.table" ) { | ||
| 613 | if ( -e $table_path && ( ( stat($table_path) )[9] > ( stat("$ramdisk_schema/$table.table") )[9] ) ) { | ||
| 614 | require File::Copy; | ||
| 615 | eval { File::Copy::copy( $table_path, "$ramdisk_schema/$table.table" ) }; | ||
| 616 | } | ||
| 617 | $target_path = "$ramdisk_schema/$table.table"; | ||
| 618 | } | ||
| 619 | |||
| 620 | if ( my $do_data = $self->do_schema($target_path) ) { | ||
| 621 | $self->normalize_blocks( $table, $do_data ); | ||
| 622 | $self->{_table}->{$table} = $do_data; | ||
| 623 | |||
| 624 | if ( $ramdisk_schema && !-e "$ramdisk_schema/$table.table" && $self->dir_exist($ramdisk_schema) ) { | ||
| 625 | require File::Copy; | ||
| 626 | eval { File::Copy::copy( $table_path, "$ramdisk_schema/$table.table" ) }; | ||
| 627 | } | ||
| 628 | if ( $do_data->{use_ramdisk} ) { | ||
| 629 | $do_data->{_ramdisk_ensured} = 1; | ||
| 630 | $self->ramdisk_ensure($table); | ||
| 631 | } | ||
| 632 | } | ||
| 633 | else { | ||
| 634 | my $schema = $self->{_table}->{$table} ||= {}; | ||
| 635 | $self->normalize_blocks( $table, $schema ); | ||
| 636 | $self->{_table}->{$table} = $schema; | ||
| 637 | if ( $schema->{use_ramdisk} && !$schema->{_ramdisk_ensured} ) { | ||
| 638 | $schema->{_ramdisk_ensured} = 1; | ||
| 639 | $self->ramdisk_ensure($table); | ||
| 640 | } | ||
| 641 | } | ||
| 642 | |||
| 643 | if ( $self->{_table}->{$table} && %{ $self->{_table}->{$table} } ) { | ||
| 644 | $self->{_table}->{$table}->{table} //= $table; | ||
| 645 | $self->{_table}->{$table}->{dbase} //= $dbase if $dbase; | ||
| 646 | |||
| 647 | $self->dbase_info($dbase) if $dbase; | ||
| 648 | |||
| 649 | return { %{ $self->{_table}->{$table} } }; | ||
| 650 | } | ||
| 651 | |||
| 652 | return {}; | ||
| 653 | } | ||
| 654 | |||
| 655 | # Checks if a table has defined schema blocks (.table file with blocks). | ||
| 656 | # Pure O(1) in-memory cache lookup. | ||
| 657 | # ------------------------------------------------ | ||
| 658 | sub has_schema { | ||
| 659 | my ( $self, $tableid ) = @_; | ||
| 660 | return 0 unless defined $tableid && length $tableid; | ||
| 661 | |||
| 662 | $self->table_info($tableid); | ||
| 663 | my $tbl_cache = $self->{_table}->{$tableid}; | ||
| 664 | |||
| 665 | return ( $tbl_cache && $tbl_cache->{_has_schema} ) ? 1 : 0; | ||
| 666 | } | ||
| 667 | |||
| 668 | # Checks if a given hash structure matches the schema defined in table_info. | ||
| 669 | # Returns 1 if table has a schema AND hash contains valid schema block keys. | ||
| 670 | # Returns 0 immediately if table has no schema OR if ANY key does not belong to schema. | ||
| 671 | # Short-circuits on the very first unmatched key with zero allocation. | ||
| 672 | # ------------------------------------------------ | ||
| 673 | sub schema_matches_hash { | ||
| 674 | my ( $self, $tableid, $hash ) = @_; | ||
| 675 | |||
| 676 | return 0 unless defined $tableid && length $tableid; | ||
| 677 | return 0 unless ref($hash) eq 'HASH' && %$hash; | ||
| 678 | |||
| 679 | my $tbl_cache = $self->{_table}->{$tableid}; | ||
| 680 | if ( !$tbl_cache || !exists $tbl_cache->{_has_schema} ) { | ||
| 681 | $self->table_info($tableid); | ||
| 682 | $tbl_cache = $self->{_table}->{$tableid}; | ||
| 683 | } | ||
| 684 | |||
| 685 | # Fast short-circuit: table without schema cannot match | ||
| 686 | return 0 unless $tbl_cache && $tbl_cache->{_has_schema}; | ||
| 687 | |||
| 688 | my $known_keys = $tbl_cache->{_block_keys}; | ||
| 689 | return 0 unless $known_keys && %$known_keys; | ||
| 690 | |||
| 691 | # Immediate rejection loop: If ANY key does NOT belong to the schema, exit immediately! | ||
| 692 | my $has_field = 0; | ||
| 693 | for my $k ( keys %$hash ) { | ||
| 694 | my $lc_k = lc($k); | ||
| 695 | next if $lc_k eq 'id' || $lc_k eq '0'; | ||
| 696 | return 0 unless $known_keys->{$lc_k}; | ||
| 697 | $has_field = 1; | ||
| 698 | } | ||
| 699 | |||
| 700 | return $has_field || exists $hash->{id} || exists $hash->{ID} || exists $hash->{0} ? 1 : 0; | ||
| 701 | } | ||
| 702 | |||
| 703 | # my $table_path = $adb->table_path($table); | ||
| 704 | # ------------------------------------------------ | ||
| 705 | sub table_path { | ||
| 706 | my ( $self, $table, $with_ext ) = @_; | ||
| 707 | |||
| 708 | $table = $self->sanitize_table($table); | ||
| 709 | return "" unless defined $table && length $table; | ||
| 710 | |||
| 711 | # return if processed earlier and still exists on disk | ||
| 712 | if ( $self->{_table}->{$table}->{_path} ) { | ||
| 713 | my $ext = $self->{db_ext} || "db"; | ||
| 714 | if ( -e ( $self->{_table}->{$table}->{_path} . ".$ext" ) ) { | ||
| 715 | return $self->{_table}->{$table}->{_path} . ($with_ext ? ".$ext" : ""); | ||
| 716 | } | ||
| 717 | delete $self->{_table}->{$table}->{_path}; | ||
| 718 | } | ||
| 719 | |||
| 720 | # load table info first | ||
| 721 | my $table_info = $self->table_info($table); | ||
| 722 | |||
| 723 | # Volatile RAM-Disk Tier 3: table lives strictly on RAM-disk | ||
| 724 | if ( $table_info && ( $table_info->{use_ramdisk} // 0 ) == 3 && $self->ramdisk_is_mounted() ) { | ||
| 725 | my $target = $self->ramdisk_path($table); | ||
| 726 | $self->{_table}->{$table}->{_path} = $target; | ||
| 727 | return $target . ( $with_ext ? ".$self->{db_ext}" : "" ); | ||
| 728 | } | ||
| 729 | |||
| 730 | # set simple mode if DATADIR directory does not exist | ||
| 731 | if ( !$self->dir_exist( $self->path('dbase_dir') ) ) { | ||
| 732 | $self->config( use_simple => 1 ); | ||
| 733 | } | ||
| 734 | |||
| 735 | # return table path if simple mode | ||
| 736 | if ( $self->config('use_simple') ) { | ||
| 737 | my $target = ( $self->path('dbase_dir') || "." ) . "/$table"; | ||
| 738 | my ($parent_dir) = $target =~ m{^(.+)[/\\][^/\\]+$}; | ||
| 739 | if ( $parent_dir && !$self->dir_exist($parent_dir) ) { | ||
| 740 | croak "[AMBERDB_FATAL] Database directory '$parent_dir' does not exist for table '$table'"; | ||
| 741 | } | ||
| 742 | $self->{_table}->{$table}->{_path} = $target; | ||
| 743 | return $target . ($with_ext ? ".$self->{db_ext}" : ""); | ||
| 744 | } | ||
| 745 | |||
| 746 | # set path value | ||
| 747 | my $dbase_dir = $self->path('dbase_dir') || "."; | ||
| 748 | |||
| 749 | my ($dbase) = ( $table =~ /^([a-z0-9]+)_/i ); | ||
| 750 | $dbase //= ""; | ||
| 751 | |||
| 752 | # if root dbase | ||
| 753 | if ( $self->{_dbase}->{$dbase}->{root} ) { | ||
| 754 | my $target = ( $self->path('dbase_dir') || "." ) . "/$table"; | ||
| 755 | my ($parent_dir) = $target =~ m{^(.+)[/\\][^/\\]+$}; | ||
| 756 | if ( $parent_dir && !$self->dir_exist($parent_dir) ) { | ||
| 757 | croak "[AMBERDB_FATAL] Database directory '$parent_dir' does not exist for table '$table'"; | ||
| 758 | } | ||
| 759 | $self->{_table}->{$table}->{_path} = $target; | ||
| 760 | return $self->{_table}->{$table}->{_path} . ($with_ext ? ".$self->{db_ext}" : ""); | ||
| 761 | } | ||
| 762 | |||
| 763 | if ( $table_info && exists $table_info->{table_dir} ) { | ||
| 764 | my $tdir = $table_info->{table_dir} // ''; | ||
| 765 | if ( $tdir ) { | ||
| 766 | $tdir =~ s{^[\\/]+|[\\/]+$}{}g; | ||
| 767 | $dbase_dir .= "/$tdir"; | ||
| 768 | } | ||
| 769 | # if length $tdir == 0 (table_dir => ''), overwrite default: keep $dbase_dir directly | ||
| 770 | } | ||
| 771 | else { | ||
| 772 | # if using year | ||
| 773 | my $yeardir; | ||
| 774 | if ( | ||
| 775 | $self->config('use_year') | ||
| 776 | and ( $self->{_dbase}->{$dbase}->{year} | ||
| 777 | or $self->{_table}->{$table}->{year} ) | ||
| 778 | ) | ||
| 779 | { | ||
| 780 | $yeardir = $self->path('year_dir') || $self->year; | ||
| 781 | } | ||
| 782 | else { | ||
| 783 | delete( $self->{_dbase}->{$dbase}->{year} ) | ||
| 784 | if ( $self->{_dbase}->{$dbase}->{year} ); | ||
| 785 | |||
| 786 | delete( $self->{_table}->{$table}->{year} ) | ||
| 787 | if ( $self->{_table}->{$table}->{year} ); | ||
| 788 | $yeardir = "table"; | ||
| 789 | } | ||
| 790 | $dbase_dir .= "/$yeardir" if $yeardir; | ||
| 791 | |||
| 792 | # if using section | ||
| 793 | if ( | ||
| 794 | $self->config('use_section') | ||
| 795 | and ( $self->{_dbase}->{$dbase}->{section} | ||
| 796 | or $self->{_table}->{$table}->{section} ) | ||
| 797 | ) { | ||
| 798 | my $section = $self->{_table}->{$table}->{section} || $self->config('section') || "center"; | ||
| 799 | $dbase_dir .= "_$section"; | ||
| 800 | } | ||
| 801 | |||
| 802 | # if using language | ||
| 803 | if ( | ||
| 804 | $self->config('use_language') and | ||
| 805 | ( $self->{_dbase}->{$dbase}->{lang} || | ||
| 806 | $self->{_table}->{$table}->{lang} ) | ||
| 807 | ) { | ||
| 808 | my $lang = $self->{_dbase}->{$dbase}->{lang} || | ||
| 809 | $self->{_table}->{$table}->{lang} // ""; | ||
| 810 | |||
| 811 | $dbase_dir .= "_$lang"; | ||
| 812 | } | ||
| 813 | } | ||
| 814 | |||
| 815 | unless ( $self->dir_exist($dbase_dir) ) { | ||
| 816 | if ( $table_info && exists $table_info->{table_dir} ) { | ||
| 817 | $self->make_path($dbase_dir); | ||
| 818 | } | ||
| 819 | else { | ||
| 820 | croak "[AMBERDB_FATAL] Database directory '$dbase_dir' does not exist for table '$table'"; | ||
| 821 | } | ||
| 822 | } | ||
| 823 | |||
| 824 | my $target = "$dbase_dir/$table"; | ||
| 825 | $self->{_table}->{$table}->{_path} = $target; | ||
| 826 | |||
| 827 | return $self->{_table}->{$table}->{_path} . ($with_ext ? ".$self->{db_ext}" : ""); | ||
| 828 | } | ||
| 829 | |||
| 830 | # my $attrs = $adb->table_attr($table); | ||
| 831 | # ------------------------------------------------ | ||
| 832 | sub table_attr { | ||
| 833 | my ( $self, $table, @args ) = @_; | ||
| 834 | |||
| 835 | $table = $self->sanitize_table($table); | ||
| 836 | return unless defined $table && length $table; | ||
| 837 | |||
| 838 | # Ensure full table schema is loaded before querying or modifying attributes | ||
| 839 | unless ( $self->{_table}->{$table} && $self->{_table}->{$table}->{_has_schema} ) { | ||
| 840 | $self->table_info($table); | ||
| 841 | } | ||
| 842 | |||
| 843 | # 1. No extra arguments: return shallow copy of table attributes | ||
| 844 | if ( !@args ) { | ||
| 845 | unless ( $self->ramdisk_is_mounted() ) { | ||
| 846 | delete $self->{_table}->{$table}->{use_ramdisk}; | ||
| 847 | delete $self->{_table}->{$table}->{use_cache}; | ||
| 848 | } | ||
| 849 | return { %{ $self->{_table}->{$table} || {} } }; | ||
| 850 | } | ||
| 851 | |||
| 852 | # 2. Single scalar argument: getter -> $adb->table_attr($table, 'use_simple') | ||
| 853 | if ( @args == 1 && !ref( $args[0] ) ) { | ||
| 854 | if ( ( $args[0] eq 'use_ramdisk' || $args[0] eq 'use_cache' ) && !$self->ramdisk_is_mounted() ) { | ||
| 855 | delete $self->{_table}->{$table}->{use_ramdisk}; | ||
| 856 | delete $self->{_table}->{$table}->{use_cache}; | ||
| 857 | return undef; | ||
| 858 | } | ||
| 859 | return $self->{_table}->{$table}->{ $args[0] }; | ||
| 860 | } | ||
| 861 | |||
| 862 | # 3. Setter: key-value list or hashref | ||
| 863 | my %attrs = ( @args == 1 && ref( $args[0] ) eq "HASH" ) ? %{ $args[0] } : @args; | ||
| 864 | my $needs_path_refresh = 0; | ||
| 865 | |||
| 866 | foreach my $key ( keys %attrs ) { | ||
| 867 | my $val = $attrs{$key}; | ||
| 868 | if ( $key eq 'use_ramdisk' ) { | ||
| 869 | $val = $self->_normalize_ramdisk_tier($val); | ||
| 870 | } | ||
| 871 | $self->{_table}->{$table}->{$key} = $val; | ||
| 872 | $needs_path_refresh = 1 if $key =~ /^(year|section|lang|table_dir|use_ramdisk)$/; | ||
| 873 | } | ||
| 874 | |||
| 875 | $self->normalize_blocks( $table, $self->{_table}->{$table} ); | ||
| 876 | |||
| 877 | my $use_ram = $self->{_table}->{$table}->{use_ramdisk} // 0; | ||
| 878 | if ($use_ram) { | ||
| 879 | $self->{_table}->{$table}->{_ramdisk_ensured} = 1; | ||
| 880 | $self->ramdisk_ensure($table); | ||
| 881 | } | ||
| 882 | |||
| 883 | if ($needs_path_refresh) { | ||
| 884 | delete $self->{_table}->{$table}->{_path}; | ||
| 885 | $self->table_path($table); | ||
| 886 | } | ||
| 887 | |||
| 888 | return $self; | ||
| 889 | } | ||
| 890 | |||
| 891 | # my $table_path = $adb->table_infset($table); | ||
| 892 | # ------------------------------------------------ | ||
| 893 | sub table_infset { | ||
| 894 | my ( $self, $table, $schema_data ) = @_; | ||
| 895 | |||
| 896 | $self->config('use_simple') and return 1; | ||
| 897 | if ( ref($schema_data) eq 'HASH' ) { | ||
| 898 | $self->normalize_blocks( $table, $schema_data ); | ||
| 899 | $self->{_table}->{$table} = $schema_data; | ||
| 900 | } | ||
| 901 | my $tbl = $self->{_table}->{$table}; | ||
| 902 | ref($tbl) eq 'HASH' or return; | ||
| 903 | |||
| 904 | my $table_path = "$table"; | ||
| 905 | $table_path =~ s/\\/\//g; | ||
| 906 | $table_path =~ s/\//-/g; | ||
| 907 | |||
| 908 | my $table_str = ""; | ||
| 909 | |||
| 910 | # Scalar keys | ||
| 911 | my @scalar_keys = qw( | ||
| 912 | name record_index keep_deleted log_owner parent_table | ||
| 913 | use_menu use_simple force use_ramdisk ramdisk_ttl use_cache use_alias | ||
| 914 | use_counter use_facet stop_word min_char | ||
| 915 | use_junk cache_ttl repeat_ids repeat_start | ||
| 916 | no_transact no_backup table_dir | ||
| 917 | ); | ||
| 918 | foreach my $key (@scalar_keys) { | ||
| 919 | next unless exists $tbl->{$key}; | ||
| 920 | next unless defined $tbl->{$key} && $tbl->{$key} ne ""; | ||
| 921 | ( my $val = $tbl->{$key} ) =~ s/"/\\"/g; | ||
| 922 | $table_str .= "\t$key => \"$val\",\n"; | ||
| 923 | } | ||
| 924 | |||
| 925 | # Array keys | ||
| 926 | my @array_keys = qw( | ||
| 927 | search_block match_block view_block facet_block | ||
| 928 | filter_block slug_block reverse suggest_block | ||
| 929 | ); | ||
| 930 | foreach my $key (@array_keys) { | ||
| 931 | next unless ref( $tbl->{$key} ) eq "ARRAY"; | ||
| 932 | my $array = join ", ", @{ $tbl->{$key} }; | ||
| 933 | next unless $array; | ||
| 934 | $table_str .= "\t$key => [ $array ],\n"; | ||
| 935 | } | ||
| 936 | |||
| 937 | # Serialize sort_block | ||
| 938 | if ( ref( $tbl->{sort_block} ) eq 'ARRAY' && @{ $tbl->{sort_block} } ) { | ||
| 939 | $table_str .= "\tsort_block => [\n"; | ||
| 940 | foreach my $sb ( @{ $tbl->{sort_block} } ) { | ||
| 941 | if ( ref($sb) eq 'HASH' ) { | ||
| 942 | $table_str .= "\t\t{ blk => $sb->{blk}, type => \"$sb->{type}\"" . ( $sb->{len} ? ", len => $sb->{len}" : "" ) . " },\n"; | ||
| 943 | } | ||
| 944 | else { | ||
| 945 | $table_str .= "\t\t$sb,\n"; | ||
| 946 | } | ||
| 947 | } | ||
| 948 | $table_str .= "\t],\n"; | ||
| 949 | } | ||
| 950 | |||
| 951 | # Serialize facet_rules | ||
| 952 | if ( ref( $tbl->{facet_rules} ) eq 'ARRAY' && @{ $tbl->{facet_rules} } ) { | ||
| 953 | my $fa = $tbl->{facet_rules}; | ||
| 954 | my $fa_val; | ||
| 955 | if ( ref( $fa->[0] ) eq 'ARRAY' ) { | ||
| 956 | my @rules_str = map { | ||
| 957 | "[" . join( ", ", map { $_ =~ /^\d+$/ ? $_ : "\"$_\"" } @$_ ) . "]" | ||
| 958 | } @$fa; | ||
| 959 | $fa_val = "[ " . join( ", ", @rules_str ) . " ]"; | ||
| 960 | } | ||
| 961 | else { | ||
| 962 | $fa_val = "[ " . join( ", ", map { $_ =~ /^\d+$/ ? $_ : "\"$_\"" } @$fa ) . " ]"; | ||
| 963 | } | ||
| 964 | $table_str .= "\tfacet_rules => $fa_val,\n"; | ||
| 965 | } | ||
| 966 | |||
| 967 | # Serialize junk_rules | ||
| 968 | if ( ref( $tbl->{junk_rules} ) eq 'ARRAY' && @{ $tbl->{junk_rules} } ) { | ||
| 969 | my $jr = $tbl->{junk_rules}; | ||
| 970 | my $jr_val; | ||
| 971 | if ( ref( $jr->[0] ) eq 'ARRAY' ) { | ||
| 972 | my @rules_str = map { | ||
| 973 | "[" . join( ", ", map { $_ =~ /^\d+$/ ? $_ : "\"$_\"" } @$_ ) . "]" | ||
| 974 | } @$jr; | ||
| 975 | $jr_val = "[ " . join( ", ", @rules_str ) . " ]"; | ||
| 976 | } | ||
| 977 | else { | ||
| 978 | $jr_val = "[ " . join( ", ", map { $_ =~ /^\d+$/ ? $_ : "\"$_\"" } @$jr ) . " ]"; | ||
| 979 | } | ||
| 980 | $table_str .= "\tjunk_rules => $jr_val,\n"; | ||
| 981 | } | ||
| 982 | |||
| 983 | # Serialize suggest_join | ||
| 984 | if ( ref( $tbl->{suggest_join} ) eq 'ARRAY' && @{ $tbl->{suggest_join} } ) { | ||
| 985 | my $sj = $tbl->{suggest_join}; | ||
| 986 | my @rules_str = map { | ||
| 987 | ref($_) eq 'ARRAY' ? "[" . join( ", ", @$_ ) . "]" : "[$_]" | ||
| 988 | } @$sj; | ||
| 989 | my $sj_val = "[ " . join( ", ", @rules_str ) . " ]"; | ||
| 990 | $table_str .= "\tsuggest_join => $sj_val,\n"; | ||
| 991 | } | ||
| 992 | |||
| 993 | # Serialize blocks in correct format | ||
| 994 | if ( ref( $tbl->{blocks} ) eq "ARRAY" && @{ $tbl->{blocks} } ) { | ||
| 995 | $table_str .= "\tblocks => [\n"; | ||
| 996 | my %seen; | ||
| 997 | foreach my $blok ( @{ $tbl->{blocks} } ) { | ||
| 998 | next unless ( defined $blok->{id} && length $blok->{id} ) && ( defined $blok->{name} && length $blok->{name} ); | ||
| 999 | next if $seen{ $blok->{id} }++; | ||
| 1000 | $table_str .= "\t\t{ id => \"$blok->{id}\","; | ||
| 1001 | $table_str .= " name => \"$blok->{name}\","; | ||
| 1002 | $blok->{type} and $table_str .= " type => \"$blok->{type}\","; | ||
| 1003 | $blok->{input} and $table_str .= " input => \"$blok->{input}\","; | ||
| 1004 | $blok->{valid} and $table_str .= " valid => \"$blok->{valid}\","; | ||
| 1005 | if ( ref( $blok->{rdbm} ) eq 'HASH' && $blok->{rdbm}{table} ) { | ||
| 1006 | my $disp = $blok->{rdbm}{display} || 0; | ||
| 1007 | $table_str .= | ||
| 1008 | " rdbm => { table => \"$blok->{rdbm}{table}\", display => $disp },"; | ||
| 1009 | } | ||
| 1010 | elsif ( defined $blok->{rdbm} && $blok->{rdbm} ne '' && !ref( $blok->{rdbm} ) ) { | ||
| 1011 | $table_str .= " rdbm => \"$blok->{rdbm}\","; | ||
| 1012 | } | ||
| 1013 | if ( ref( $blok->{extend} ) eq 'HASH' && $blok->{extend}{table} ) { | ||
| 1014 | my $join_col = $blok->{extend}{join} || 'id'; | ||
| 1015 | $table_str .= | ||
| 1016 | " extend => { table => \"$blok->{extend}{table}\", join => \"$join_col\" },"; | ||
| 1017 | } | ||
| 1018 | $blok->{option} and | ||
| 1019 | $table_str .= " option => \"$blok->{option}\","; | ||
| 1020 | $table_str .= " },\n"; | ||
| 1021 | } | ||
| 1022 | $table_str .= "\t],\n"; | ||
| 1023 | } | ||
| 1024 | |||
| 1025 | if ($table_str) { | ||
| 1026 | my $schema_dir = $self->path('schema_dir') || ( $self->path('dbase_dir') ? $self->path('dbase_dir') . "/schema" : "schema" ); | ||
| 1027 | open my $YZ, ">:encoding(UTF-8)", "$schema_dir/$table_path.table" | ||
| 1028 | or do { | ||
| 1029 | cluck "[DB_SCHEMA] Could not write schema $schema_dir/$table_path.table: $!\n"; | ||
| 1030 | return; | ||
| 1031 | }; | ||
| 1032 | print $YZ "{\n"; | ||
| 1033 | print $YZ $table_str; | ||
| 1034 | print $YZ "}\n"; | ||
| 1035 | close $YZ; | ||
| 1036 | } | ||
| 1037 | |||
| 1038 | return 1; | ||
| 1039 | } | ||
| 1040 | |||
| 1041 | # my $norm_info = $adb->norm_info($table_id); | ||
| 1042 | # ------------------------------------------------ | ||
| 1043 | sub norm_info { | ||
| 1044 | my ( $self, $arg ) = @_; | ||
| 1045 | |||
| 1046 | return {} unless $arg; | ||
| 1047 | return {} if $self->config('use_simple'); | ||
| 1048 | |||
| 1049 | my ( $table, $norm_path ) = $self->schema_arg( $arg, "norm" ); | ||
| 1050 | $table && $norm_path or return {}; | ||
| 1051 | |||
| 1052 | $self->{_norm} ||= {}; | ||
| 1053 | if ( $self->{_norm}->{$table} && %{ $self->{_norm}->{$table} } ) { | ||
| 1054 | return { %{ $self->{_norm}->{$table} } }; | ||
| 1055 | } | ||
| 1056 | |||
| 1057 | if ( my $do_data = $self->do_schema($norm_path) ) { | ||
| 1058 | $self->{_norm}->{$table} = $do_data; | ||
| 1059 | return $do_data; | ||
| 1060 | } | ||
| 1061 | return {}; | ||
| 1062 | } | ||
| 1063 | |||
| 1064 | # my $norm_hash = $adb->field_normalize($table_id, $raw_string, [$opts]); | ||
| 1065 | # ------------------------------------------------ | ||
| 1066 | sub field_normalize { | ||
| 1067 | my ( $self, $table_id, $raw, $opts ) = @_; | ||
| 1068 | return {} unless defined $raw && $raw ne ''; | ||
| 1069 | |||
| 1070 | my $norm_info = $self->norm_info($table_id); | ||
| 1071 | return { raw_features => $raw } unless $norm_info && %$norm_info; | ||
| 1072 | |||
| 1073 | my $str = "$raw"; | ||
| 1074 | $str =~ s/<br\s*\/?>/ /gi; | ||
| 1075 | $str =~ s/\\n/ /g; | ||
| 1076 | $str =~ s/\\r/ /g; | ||
| 1077 | $str =~ s/\\t/ /g; | ||
| 1078 | |||
| 1079 | # Preprocess | ||
| 1080 | if ( $norm_info->{preprocess} ) { | ||
| 1081 | for my $p ( @{ $norm_info->{preprocess} } ) { | ||
| 1082 | my ( $pat, $repl ) = @$p; | ||
| 1083 | if ( $repl eq q{$1 $2} ) { | ||
| 1084 | $str =~ s/$pat/$1 $2/g; | ||
| 1085 | } | ||
| 1086 | elsif ( $repl eq q{. $1} ) { | ||
| 1087 | $str =~ s/$pat/. $1/g; | ||
| 1088 | } | ||
| 1089 | elsif ( $repl eq q{$1 $2 Sayfa} ) { | ||
| 1090 | $str =~ s/$pat/$1 $2 Sayfa/gi; | ||
| 1091 | } | ||
| 1092 | else { | ||
| 1093 | $str =~ s/$pat/$repl/g; | ||
| 1094 | } | ||
| 1095 | } | ||
| 1096 | } | ||
| 1097 | |||
| 1098 | my %res; | ||
| 1099 | if ( $norm_info->{rules} ) { | ||
| 1100 | for my $r ( @{ $norm_info->{rules} } ) { | ||
| 1101 | my $field = $r->{field}; | ||
| 1102 | |||
| 1103 | if ( $r->{map} ) { | ||
| 1104 | for my $m ( @{ $r->{map} } ) { | ||
| 1105 | my ( $pat, $target ) = @$m; | ||
| 1106 | if ( $str =~ $pat ) { | ||
| 1107 | $res{$field} = $target; | ||
| 1108 | last; | ||
| 1109 | } | ||
| 1110 | } | ||
| 1111 | } | ||
| 1112 | elsif ( $r->{match_all} ) { | ||
| 1113 | my @matches = ( $str =~ /$r->{match_all}/g ); | ||
| 1114 | if (@matches) { | ||
| 1115 | if ( ( $r->{select} // 'max' ) eq 'max' ) { | ||
| 1116 | my ($m) = sort { $b <=> $a } @matches; | ||
| 1117 | $res{$field} = $m + 0; | ||
| 1118 | } | ||
| 1119 | elsif ( $r->{select} eq 'min' ) { | ||
| 1120 | my ($m) = sort { $a <=> $b } @matches; | ||
| 1121 | $res{$field} = $m + 0; | ||
| 1122 | } | ||
| 1123 | } | ||
| 1124 | } | ||
| 1125 | elsif ( $r->{match} ) { | ||
| 1126 | if ( $str =~ $r->{match} ) { | ||
| 1127 | if ( $r->{format} && ref( $r->{format} ) eq 'CODE' ) { | ||
| 1128 | $res{$field} = $r->{format}->( $1, $2, $3 ); | ||
| 1129 | } | ||
| 1130 | elsif ( ( $r->{type} // '' ) eq 'integer' ) { | ||
| 1131 | $res{$field} = $1 + 0; | ||
| 1132 | } | ||
| 1133 | elsif ( ( $r->{type} // '' ) eq 'titlecase' ) { | ||
| 1134 | $res{$field} = ucfirst( lc($1) ); | ||
| 1135 | } | ||
| 1136 | elsif ( ( $r->{type} // '' ) eq 'trim' ) { | ||
| 1137 | my $val = $1; | ||
| 1138 | $val =~ s/^\s+|\s+$//g; | ||
| 1139 | $res{$field} = $val; | ||
| 1140 | } | ||
| 1141 | else { | ||
| 1142 | $res{$field} = $1; | ||
| 1143 | } | ||
| 1144 | } | ||
| 1145 | } | ||
| 1146 | } | ||
| 1147 | } | ||
| 1148 | |||
| 1149 | $res{raw_features} = $raw; | ||
| 1150 | |||
| 1151 | # If source_block or target table schema mapping is requested | ||
| 1152 | if ( $opts && $opts->{as_array} ) { | ||
| 1153 | my $table_info = $self->table_info($table_id); | ||
| 1154 | if ( $table_info && $table_info->{blocks} ) { | ||
| 1155 | my @rec; | ||
| 1156 | my $blocks = $table_info->{blocks}; | ||
| 1157 | for my $b (@$blocks) { | ||
| 1158 | my $fid = $b->{id}; | ||
| 1159 | push @rec, ( $res{$fid} // '' ); | ||
| 1160 | } | ||
| 1161 | return \@rec; | ||
| 1162 | } | ||
| 1163 | } | ||
| 1164 | |||
| 1165 | return \%res; | ||
| 1166 | } | ||
| 1167 | |||
| 1168 | 1; |