Update prompts & parsing

stages-tlr/model-provider/src/main/java/edu/kit/kastel/mcse/ardoco/tlr/models/informants/LLMArchitecturePrompt.java

@@ -10,7 +10,12 @@ public enum LLMArchitecturePrompt {
 
                     %s
                     """,
-            "Now provide a list that only covers the component names. Omit common prefixes and suffixes in the names."), //
+            """
+                    Now provide a list that only covers the component names in camel case. Omit common prefixes and suffixes.
+                    Output format:
+                    - Name1
+                    - Name2
+                    """), //
     CODE_ONLY_V1(
             """
                     You get the Packages of a software project. Your task is to summarize the Packages w.r.t. the high-level architecture of the system. Try to identify possible components.
@@ -19,10 +24,18 @@ public enum LLMArchitecturePrompt {
 
                     %s
                     """,
-            "Now provide a list that only covers the component names. Omit common prefixes and suffixes in the names."), //
+            """
+                    Now provide a list that only covers the component names. Omit common prefixes and suffixes in the names in camel case.
+                    Output format:
+                    - Name1
+                    - Name2
+                    """), //
     AGGREGATION_V1("""
             You get a list of possible component names. Your task is to aggregate the list and remove duplicates.
-            Also filter out component names that are very generic. Do not repeat what you filtered out. Only provide the final enumeration.
+            Also filter out component names that are very generic. Provide only the final component names in camel case.
+            Output format:
+            - Name1
+            - Name2
 
             Possible component names:
 

stages-tlr/model-provider/src/main/java/edu/kit/kastel/mcse/ardoco/tlr/models/informants/LLMArchitectureProviderInformant.java

@@ -81,9 +81,8 @@ protected void process() {
 
         // Remove any not letter characters
         componentNames = componentNames.stream()
-                .map(it -> it.replaceAll("[^a-zA-Z0-9 \\-_]", "").replaceAll("\\s+", " ").trim())
+                // .map(it -> it.replaceAll("[^a-zA-Z0-9 \\-_]", "").replaceAll("\\s+", " ").trim())
                 .map(it -> it.replace("Components", "").replace("Component", "").trim())
-                .map(it -> it.replace(" ", ""))
                 .filter(it -> !it.isBlank())
                 .distinct()
                 .sorted()
@@ -162,31 +161,19 @@ private void parseComponentNames(String response, List<String> componentNames) {
             }
             line = line.trim();
 
-            // Version 5: 1. Name (NotImportant) or 2. Name (SomeString)
-            if (line.matches("^\\d+\\.\\s+.+\\s*\\(.*\\)$")) {
-                componentNames.add(line.split("\\.\\s+")[1].split("\\s*\\(.*\\)")[0]);
-            }
-            // Version 1: 1. **Name** or 2. **Name**
-            else if (line.matches("^\\d+\\.\\s*\\*\\*.*\\*\\*$")) {
-                componentNames.add(line.split("\\*\\*")[1]);
-            }
-            // Version 2: 1. Name or 2. Name
-            else if (line.matches("^\\d+\\.\\s+.*$")) {
-                componentNames.add(line.split("\\d+\\.\\s+")[1]);
-            }
-            // Version 3: - **Name**
-            else if (line.matches("^([-*])\\s+\\*\\*.*\\*\\*$")) {
-                componentNames.add(line.split("\\*\\*")[1]);
-            }
-            // Version 4: - Name
-            else if (line.matches("^([-*])\\s+.*$")) {
-                componentNames.add(line.split("([-*])\\s+")[1]);
-            }
-            // Version 5: Name, Name2, Name3, (at least 4 names .. otherwise we match to many things)
-            else if (line.matches("(.*,\\s+){3,}.*")) {
-                componentNames.addAll(List.of(line.split(",\\s+")));
-            } else {
-                logger.warn("Could not parse line: {}", line);
+            if (line.startsWith("-")) {
+                // Defined Format "- Name1"
+                var name = line.substring(1).trim();
+                componentNames.add(name);
+            } /* else if (Character.isDigit(line.charAt(0)) && line.contains(".")) {
+                // Fallback Format: 1. Name
+                var name = line.split("\\.", 2)[1].trim();
+                // We defined camel case ... so all after the space might be additional information
+                if (name.contains(" "))
+                    name = name.split(" ", 2)[0].trim();
+                componentNames.add(name);
+              }*/ else {
+                logger.warn("Could not parse component name: {}", line);
             }
         }
     }

stages-tlr/model-provider/src/main/java/edu/kit/kastel/mcse/ardoco/tlr/models/informants/LargeLanguageModel.java

@@ -69,7 +69,13 @@ private static ChatLanguageModel createOpenAiModel(String model) {
         if (apiKey == null || orgId == null) {
             throw new IllegalArgumentException("OPENAI_API_KEY and OPENAI_ORG_ID must be set as environment variables");
         }
-        return new OpenAiChatModel.OpenAiChatModelBuilder().modelName(model).apiKey(apiKey).organizationId(orgId).seed(SEED).temperature(0.0).build();
+        return new OpenAiChatModel.OpenAiChatModelBuilder().modelName(model)
+                .apiKey(apiKey)
+                .organizationId(orgId)
+                .seed(SEED)
+                .temperature(0.0)
+                .timeout(Duration.ofMinutes(10))
+                .build();
     }
 
     private static ChatLanguageModel createOllamaModel(String model) {