Akron | 57ee558 | 2025-05-21 15:25:13 +0200 | [diff] [blame] | 1 | package config |
| 2 | |
| 3 | import ( |
| 4 | "fmt" |
| 5 | "os" |
| 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/parser" |
Akron | 57ee558 | 2025-05-21 15:25:13 +0200 | [diff] [blame] | 9 | "gopkg.in/yaml.v3" |
| 10 | ) |
| 11 | |
Akron | 06d21f0 | 2025-06-04 14:36:07 +0200 | [diff] [blame^] | 12 | const ( |
| 13 | defaultServer = "https://korap.ids-mannheim.de/" |
| 14 | defaultSDK = "https://korap.ids-mannheim.de/js/korap-plugin-latest.js" |
| 15 | ) |
| 16 | |
Akron | 57ee558 | 2025-05-21 15:25:13 +0200 | [diff] [blame] | 17 | // MappingRule represents a single mapping rule in the configuration |
| 18 | type MappingRule string |
| 19 | |
| 20 | // MappingList represents a list of mapping rules with metadata |
| 21 | type MappingList struct { |
| 22 | ID string `yaml:"id"` |
| 23 | FoundryA string `yaml:"foundryA,omitempty"` |
| 24 | LayerA string `yaml:"layerA,omitempty"` |
| 25 | FoundryB string `yaml:"foundryB,omitempty"` |
| 26 | LayerB string `yaml:"layerB,omitempty"` |
| 27 | Mappings []MappingRule `yaml:"mappings"` |
| 28 | } |
| 29 | |
Akron | 06d21f0 | 2025-06-04 14:36:07 +0200 | [diff] [blame^] | 30 | // MappingConfig represents the root configuration containing multiple mapping lists |
| 31 | type MappingConfig struct { |
| 32 | SDK string `yaml:"sdk,omitempty"` |
| 33 | Server string `yaml:"server,omitempty"` |
| 34 | Lists []MappingList `yaml:"lists,omitempty"` |
Akron | 57ee558 | 2025-05-21 15:25:13 +0200 | [diff] [blame] | 35 | } |
| 36 | |
| 37 | // LoadConfig loads a YAML configuration file and returns a Config object |
Akron | 06d21f0 | 2025-06-04 14:36:07 +0200 | [diff] [blame^] | 38 | func LoadConfig(filename string) (*MappingConfig, error) { |
Akron | 57ee558 | 2025-05-21 15:25:13 +0200 | [diff] [blame] | 39 | data, err := os.ReadFile(filename) |
| 40 | if err != nil { |
| 41 | return nil, fmt.Errorf("failed to read config file: %w", err) |
| 42 | } |
| 43 | |
Akron | a5d8814 | 2025-05-22 14:42:09 +0200 | [diff] [blame] | 44 | // Check for empty file |
| 45 | if len(data) == 0 { |
| 46 | return nil, fmt.Errorf("EOF: config file is empty") |
| 47 | } |
| 48 | |
Akron | 06d21f0 | 2025-06-04 14:36:07 +0200 | [diff] [blame^] | 49 | // Try to unmarshal as new format first (object with optional sdk/server and lists) |
| 50 | var config MappingConfig |
| 51 | if err := yaml.Unmarshal(data, &config); err == nil && len(config.Lists) > 0 { |
| 52 | // Successfully parsed as new format with lists field |
| 53 | if err := validateMappingLists(config.Lists); err != nil { |
| 54 | return nil, err |
| 55 | } |
| 56 | // Apply defaults if not specified |
| 57 | applyDefaults(&config) |
| 58 | return &config, nil |
| 59 | } |
| 60 | |
| 61 | // Fall back to old format (direct list) |
Akron | 57ee558 | 2025-05-21 15:25:13 +0200 | [diff] [blame] | 62 | var lists []MappingList |
| 63 | if err := yaml.Unmarshal(data, &lists); err != nil { |
| 64 | return nil, fmt.Errorf("failed to parse YAML: %w", err) |
| 65 | } |
| 66 | |
Akron | 06d21f0 | 2025-06-04 14:36:07 +0200 | [diff] [blame^] | 67 | if err := validateMappingLists(lists); err != nil { |
| 68 | return nil, err |
| 69 | } |
| 70 | |
| 71 | config = MappingConfig{Lists: lists} |
| 72 | // Apply defaults if not specified |
| 73 | applyDefaults(&config) |
| 74 | return &config, nil |
| 75 | } |
| 76 | |
| 77 | // applyDefaults sets default values for SDK and Server if they are empty |
| 78 | func applyDefaults(config *MappingConfig) { |
| 79 | if config.SDK == "" { |
| 80 | config.SDK = defaultSDK |
| 81 | } |
| 82 | if config.Server == "" { |
| 83 | config.Server = defaultServer |
| 84 | } |
| 85 | } |
| 86 | |
| 87 | // validateMappingLists validates a slice of mapping lists |
| 88 | func validateMappingLists(lists []MappingList) error { |
Akron | 57ee558 | 2025-05-21 15:25:13 +0200 | [diff] [blame] | 89 | // Validate the configuration |
Akron | a5d8814 | 2025-05-22 14:42:09 +0200 | [diff] [blame] | 90 | seenIDs := make(map[string]bool) |
Akron | 57ee558 | 2025-05-21 15:25:13 +0200 | [diff] [blame] | 91 | for i, list := range lists { |
| 92 | if list.ID == "" { |
Akron | 06d21f0 | 2025-06-04 14:36:07 +0200 | [diff] [blame^] | 93 | return fmt.Errorf("mapping list at index %d is missing an ID", i) |
Akron | 57ee558 | 2025-05-21 15:25:13 +0200 | [diff] [blame] | 94 | } |
Akron | a5d8814 | 2025-05-22 14:42:09 +0200 | [diff] [blame] | 95 | |
| 96 | // Check for duplicate IDs |
| 97 | if seenIDs[list.ID] { |
Akron | 06d21f0 | 2025-06-04 14:36:07 +0200 | [diff] [blame^] | 98 | return fmt.Errorf("duplicate mapping list ID found: %s", list.ID) |
Akron | a5d8814 | 2025-05-22 14:42:09 +0200 | [diff] [blame] | 99 | } |
| 100 | seenIDs[list.ID] = true |
| 101 | |
Akron | 57ee558 | 2025-05-21 15:25:13 +0200 | [diff] [blame] | 102 | if len(list.Mappings) == 0 { |
Akron | 06d21f0 | 2025-06-04 14:36:07 +0200 | [diff] [blame^] | 103 | return fmt.Errorf("mapping list '%s' has no mapping rules", list.ID) |
Akron | 57ee558 | 2025-05-21 15:25:13 +0200 | [diff] [blame] | 104 | } |
| 105 | |
| 106 | // Validate each mapping rule |
| 107 | for j, rule := range list.Mappings { |
| 108 | if rule == "" { |
Akron | 06d21f0 | 2025-06-04 14:36:07 +0200 | [diff] [blame^] | 109 | return fmt.Errorf("mapping list '%s' rule at index %d is empty", list.ID, j) |
Akron | 57ee558 | 2025-05-21 15:25:13 +0200 | [diff] [blame] | 110 | } |
| 111 | } |
| 112 | } |
Akron | 06d21f0 | 2025-06-04 14:36:07 +0200 | [diff] [blame^] | 113 | return nil |
Akron | 57ee558 | 2025-05-21 15:25:13 +0200 | [diff] [blame] | 114 | } |
| 115 | |
| 116 | // ParseMappings parses all mapping rules in a list and returns a slice of parsed rules |
| 117 | func (list *MappingList) ParseMappings() ([]*parser.MappingResult, error) { |
| 118 | // Create a grammar parser with the list's default foundries and layers |
| 119 | grammarParser, err := parser.NewGrammarParser("", "") |
| 120 | if err != nil { |
| 121 | return nil, fmt.Errorf("failed to create grammar parser: %w", err) |
| 122 | } |
| 123 | |
| 124 | results := make([]*parser.MappingResult, len(list.Mappings)) |
| 125 | for i, rule := range list.Mappings { |
Akron | a5d8814 | 2025-05-22 14:42:09 +0200 | [diff] [blame] | 126 | // Check for empty rules first |
| 127 | if rule == "" { |
| 128 | return nil, fmt.Errorf("empty mapping rule at index %d in list '%s'", i, list.ID) |
| 129 | } |
| 130 | |
Akron | 57ee558 | 2025-05-21 15:25:13 +0200 | [diff] [blame] | 131 | // Parse the mapping rule |
| 132 | result, err := grammarParser.ParseMapping(string(rule)) |
| 133 | if err != nil { |
| 134 | return nil, fmt.Errorf("failed to parse mapping rule %d in list '%s': %w", i, list.ID, err) |
| 135 | } |
| 136 | |
| 137 | // Apply default foundries and layers if not specified in the rule |
| 138 | if list.FoundryA != "" { |
| 139 | applyDefaultFoundryAndLayer(result.Upper.Wrap, list.FoundryA, list.LayerA) |
| 140 | } |
| 141 | if list.FoundryB != "" { |
| 142 | applyDefaultFoundryAndLayer(result.Lower.Wrap, list.FoundryB, list.LayerB) |
| 143 | } |
| 144 | |
| 145 | results[i] = result |
| 146 | } |
| 147 | |
| 148 | return results, nil |
| 149 | } |
| 150 | |
| 151 | // applyDefaultFoundryAndLayer recursively applies default foundry and layer to terms that don't have them specified |
| 152 | func applyDefaultFoundryAndLayer(node ast.Node, defaultFoundry, defaultLayer string) { |
| 153 | switch n := node.(type) { |
| 154 | case *ast.Term: |
| 155 | if n.Foundry == "" { |
| 156 | n.Foundry = defaultFoundry |
| 157 | } |
| 158 | if n.Layer == "" { |
| 159 | n.Layer = defaultLayer |
| 160 | } |
| 161 | case *ast.TermGroup: |
| 162 | for _, op := range n.Operands { |
| 163 | applyDefaultFoundryAndLayer(op, defaultFoundry, defaultLayer) |
| 164 | } |
| 165 | } |
| 166 | } |