Akron | 32d53de | 2025-05-22 13:45:32 +0200 | [diff] [blame] | 1 | package mapper |
| 2 | |
| 3 | import ( |
| 4 | "encoding/json" |
| 5 | "fmt" |
| 6 | |
Akron | fa55bb2 | 2025-05-26 15:10:42 +0200 | [diff] [blame] | 7 | "github.com/KorAP/KoralPipe-TermMapper/ast" |
| 8 | "github.com/KorAP/KoralPipe-TermMapper/config" |
| 9 | "github.com/KorAP/KoralPipe-TermMapper/matcher" |
| 10 | "github.com/KorAP/KoralPipe-TermMapper/parser" |
Akron | 32d53de | 2025-05-22 13:45:32 +0200 | [diff] [blame] | 11 | ) |
| 12 | |
| 13 | // Direction represents the mapping direction (A to B or B to A) |
| 14 | type Direction string |
| 15 | |
| 16 | const ( |
| 17 | AtoB Direction = "atob" |
| 18 | BtoA Direction = "btoa" |
| 19 | ) |
| 20 | |
| 21 | // Mapper handles the application of mapping rules to JSON objects |
| 22 | type Mapper struct { |
| 23 | mappingLists map[string]*config.MappingList |
| 24 | parsedRules map[string][]*parser.MappingResult |
| 25 | } |
| 26 | |
| 27 | // NewMapper creates a new Mapper instance |
| 28 | func NewMapper(configFiles ...string) (*Mapper, error) { |
| 29 | m := &Mapper{ |
| 30 | mappingLists: make(map[string]*config.MappingList), |
| 31 | parsedRules: make(map[string][]*parser.MappingResult), |
| 32 | } |
| 33 | |
| 34 | // Load and parse all config files |
| 35 | for _, file := range configFiles { |
| 36 | cfg, err := config.LoadConfig(file) |
| 37 | if err != nil { |
| 38 | return nil, fmt.Errorf("failed to load config from %s: %w", file, err) |
| 39 | } |
| 40 | |
| 41 | // Store mapping lists by ID |
| 42 | for _, list := range cfg.Lists { |
| 43 | if _, exists := m.mappingLists[list.ID]; exists { |
| 44 | return nil, fmt.Errorf("duplicate mapping list ID found: %s", list.ID) |
| 45 | } |
| 46 | |
| 47 | // Create a copy of the list to store |
| 48 | listCopy := list |
| 49 | m.mappingLists[list.ID] = &listCopy |
| 50 | |
| 51 | // Parse the rules immediately |
| 52 | parsedRules, err := list.ParseMappings() |
| 53 | if err != nil { |
| 54 | return nil, fmt.Errorf("failed to parse mappings for list %s: %w", list.ID, err) |
| 55 | } |
| 56 | m.parsedRules[list.ID] = parsedRules |
| 57 | } |
| 58 | } |
| 59 | |
| 60 | return m, nil |
| 61 | } |
| 62 | |
| 63 | // MappingOptions contains the options for applying mappings |
| 64 | type MappingOptions struct { |
| 65 | FoundryA string |
| 66 | LayerA string |
| 67 | FoundryB string |
| 68 | LayerB string |
| 69 | Direction Direction |
| 70 | } |
| 71 | |
| 72 | // ApplyMappings applies the specified mapping rules to a JSON object |
Akron | d5850f8 | 2025-05-23 16:44:44 +0200 | [diff] [blame] | 73 | func (m *Mapper) ApplyMappings(mappingID string, opts MappingOptions, jsonData any) (any, error) { |
Akron | 32d53de | 2025-05-22 13:45:32 +0200 | [diff] [blame] | 74 | // Validate mapping ID |
| 75 | if _, exists := m.mappingLists[mappingID]; !exists { |
| 76 | return nil, fmt.Errorf("mapping list with ID %s not found", mappingID) |
| 77 | } |
| 78 | |
| 79 | // Validate direction |
| 80 | if opts.Direction != AtoB && opts.Direction != BtoA { |
| 81 | return nil, fmt.Errorf("invalid direction: %s", opts.Direction) |
| 82 | } |
| 83 | |
| 84 | // Get the parsed rules |
| 85 | rules := m.parsedRules[mappingID] |
| 86 | |
| 87 | // Convert input JSON to AST |
| 88 | jsonBytes, err := json.Marshal(jsonData) |
| 89 | if err != nil { |
| 90 | return nil, fmt.Errorf("failed to marshal input JSON: %w", err) |
| 91 | } |
| 92 | |
| 93 | node, err := parser.ParseJSON(jsonBytes) |
| 94 | if err != nil { |
| 95 | return nil, fmt.Errorf("failed to parse JSON into AST: %w", err) |
| 96 | } |
| 97 | |
Akron | d5850f8 | 2025-05-23 16:44:44 +0200 | [diff] [blame] | 98 | // Store whether the input was a Token |
| 99 | isToken := false |
| 100 | var tokenWrap ast.Node |
Akron | 32d53de | 2025-05-22 13:45:32 +0200 | [diff] [blame] | 101 | if token, ok := node.(*ast.Token); ok { |
Akron | d5850f8 | 2025-05-23 16:44:44 +0200 | [diff] [blame] | 102 | isToken = true |
| 103 | tokenWrap = token.Wrap |
| 104 | node = tokenWrap |
Akron | 32d53de | 2025-05-22 13:45:32 +0200 | [diff] [blame] | 105 | } |
| 106 | |
| 107 | // Apply each rule to the AST |
| 108 | for _, rule := range rules { |
| 109 | // Create pattern and replacement based on direction |
| 110 | var pattern, replacement ast.Node |
| 111 | if opts.Direction == AtoB { |
| 112 | pattern = rule.Upper |
| 113 | replacement = rule.Lower |
| 114 | } else { |
| 115 | pattern = rule.Lower |
| 116 | replacement = rule.Upper |
| 117 | } |
| 118 | |
| 119 | // Extract the inner nodes from the pattern and replacement tokens |
| 120 | if token, ok := pattern.(*ast.Token); ok { |
| 121 | pattern = token.Wrap |
| 122 | } |
| 123 | if token, ok := replacement.(*ast.Token); ok { |
| 124 | replacement = token.Wrap |
| 125 | } |
| 126 | |
| 127 | // Apply foundry and layer overrides |
| 128 | if opts.Direction == AtoB { |
| 129 | applyFoundryAndLayerOverrides(pattern, opts.FoundryA, opts.LayerA) |
| 130 | applyFoundryAndLayerOverrides(replacement, opts.FoundryB, opts.LayerB) |
| 131 | } else { |
| 132 | applyFoundryAndLayerOverrides(pattern, opts.FoundryB, opts.LayerB) |
| 133 | applyFoundryAndLayerOverrides(replacement, opts.FoundryA, opts.LayerA) |
| 134 | } |
| 135 | |
| 136 | // Create matcher and apply replacement |
Akron | d5850f8 | 2025-05-23 16:44:44 +0200 | [diff] [blame] | 137 | m, err := matcher.NewMatcher(ast.Pattern{Root: pattern}, ast.Replacement{Root: replacement}) |
| 138 | if err != nil { |
| 139 | return nil, fmt.Errorf("failed to create matcher: %w", err) |
| 140 | } |
Akron | 32d53de | 2025-05-22 13:45:32 +0200 | [diff] [blame] | 141 | node = m.Replace(node) |
| 142 | } |
| 143 | |
Akron | d5850f8 | 2025-05-23 16:44:44 +0200 | [diff] [blame] | 144 | // Wrap the result in a token if the input was a token |
| 145 | var result ast.Node |
| 146 | if isToken { |
| 147 | result = &ast.Token{Wrap: node} |
| 148 | } else { |
| 149 | result = node |
| 150 | } |
Akron | 32d53de | 2025-05-22 13:45:32 +0200 | [diff] [blame] | 151 | |
| 152 | // Convert AST back to JSON |
| 153 | resultBytes, err := parser.SerializeToJSON(result) |
| 154 | if err != nil { |
| 155 | return nil, fmt.Errorf("failed to serialize AST to JSON: %w", err) |
| 156 | } |
| 157 | |
| 158 | // Parse the JSON string back into an interface{} |
| 159 | var resultData interface{} |
| 160 | if err := json.Unmarshal(resultBytes, &resultData); err != nil { |
| 161 | return nil, fmt.Errorf("failed to parse result JSON: %w", err) |
| 162 | } |
| 163 | |
| 164 | return resultData, nil |
| 165 | } |
| 166 | |
| 167 | // applyFoundryAndLayerOverrides recursively applies foundry and layer overrides to terms |
| 168 | func applyFoundryAndLayerOverrides(node ast.Node, foundry, layer string) { |
| 169 | if node == nil { |
| 170 | return |
| 171 | } |
| 172 | |
| 173 | switch n := node.(type) { |
| 174 | case *ast.Term: |
| 175 | if foundry != "" { |
| 176 | n.Foundry = foundry |
| 177 | } |
| 178 | if layer != "" { |
| 179 | n.Layer = layer |
| 180 | } |
| 181 | case *ast.TermGroup: |
| 182 | for _, op := range n.Operands { |
| 183 | applyFoundryAndLayerOverrides(op, foundry, layer) |
| 184 | } |
| 185 | case *ast.Token: |
| 186 | if n.Wrap != nil { |
| 187 | applyFoundryAndLayerOverrides(n.Wrap, foundry, layer) |
| 188 | } |
| 189 | case *ast.CatchallNode: |
| 190 | if n.Wrap != nil { |
| 191 | applyFoundryAndLayerOverrides(n.Wrap, foundry, layer) |
| 192 | } |
| 193 | for _, op := range n.Operands { |
| 194 | applyFoundryAndLayerOverrides(op, foundry, layer) |
| 195 | } |
| 196 | } |
| 197 | } |