Develop Q&A on Doc scenario

src/main/java/com/olympus/hermione/services/ScenarioExecutionService.java

@@ -49,6 +49,8 @@ import com.olympus.hermione.stepSolvers.AdvancedAIPromptSolver;
 import com.olympus.hermione.stepSolvers.SummarizeDocSolver;
 import com.olympus.hermione.stepSolvers.BasicAIPromptSolver;
 import com.olympus.hermione.stepSolvers.BasicQueryRagSolver;
+import com.olympus.hermione.stepSolvers.DeleteDocTempSolver;
+import com.olympus.hermione.stepSolvers.EmbeddingDocTempSolver;
 import com.olympus.hermione.stepSolvers.QueryNeo4JSolver;
 import com.olympus.hermione.stepSolvers.SourceCodeRagSolver;
 import com.olympus.hermione.stepSolvers.StepSolver;
@@ -267,6 +269,12 @@ public class ScenarioExecutionService {
             case "OLYMPUS_QUERY_AI":
                 solver = new OlynmpusChatClientSolver();
                 break;
+            case "EMBED_TEMPORARY_DOC":
+                solver = new EmbeddingDocTempSolver();
+                break;
+            case "DELETE_TEMPORARY_DOC":
+                solver = new DeleteDocTempSolver();
+                break;
             default:
                 break;
         }

src/main/java/com/olympus/hermione/stepSolvers/DeleteDocTempSolver.java

@@ -0,0 +1,75 @@
+package com.olympus.hermione.stepSolvers;
+
+import ch.qos.logback.classic.Logger;
+import com.olympus.hermione.models.ScenarioExecution;
+import com.olympus.hermione.utility.AttributeParser;
+import java.util.List;
+import org.slf4j.LoggerFactory;
+import org.springframework.ai.document.Document;
+import org.springframework.ai.vectorstore.SearchRequest;
+
+public class DeleteDocTempSolver extends StepSolver {
+    
+    private String rag_filter;
+    private int topk;
+    private double threshold;
+    private String query;
+
+
+    Logger logger = (Logger) LoggerFactory.getLogger(BasicQueryRagSolver.class);
+
+    private void loadParameters(){
+        logger.info("Loading parameters");
+        this.scenarioExecution.getExecSharedMap().put("scenario_execution_id", this.scenarioExecution.getId());
+
+        AttributeParser attributeParser = new AttributeParser(this.scenarioExecution);
+
+        // this.scenario_execution_id = attributeParser.parse((String) this.scenarioExecution.getId());
+        
+        this.rag_filter = attributeParser.parse((String) this.step.getAttributes().get("rag_filter"));
+
+        if(this.step.getAttributes().containsKey("rag_query")){
+            this.query = (String) this.step.getAttributes().get("rag_query");
+        }else{
+            this.query = "*";
+        }
+
+        if(this.step.getAttributes().containsKey("rag_topk")){
+            this.topk = (int) this.step.getAttributes().get("rag_topk");
+        }else{
+            this.topk = 1000;
+        }
+
+        if(this.step.getAttributes().containsKey("rag_threshold")){
+            this.threshold = (double) this.step.getAttributes().get("rag_threshold");
+        }else{
+           this.threshold = 0.0;
+        }
+
+    }
+
+    @Override
+    public ScenarioExecution solveStep(){
+    
+        System.out.println("Solving step: " + this.step.getName());
+        
+        this.scenarioExecution.setCurrentStepId(this.step.getStepId());
+
+        loadParameters();
+        SearchRequest searchRequest = SearchRequest.defaults()
+                                        .withQuery(this.query)
+                                        .withTopK(this.topk)
+                                        .withSimilarityThreshold(this.threshold)
+                                        .withFilterExpression(this.rag_filter);
+
+
+        List<Document> docs  = vectorStore.similaritySearch(searchRequest);
+        List<String> ids = docs.stream().map(Document::getId).toList();
+        vectorStore.delete(ids);
+
+        this.scenarioExecution.setNextStepId(this.step.getNextStepId());
+
+        return this.scenarioExecution;
+    }
+
+}

src/main/java/com/olympus/hermione/stepSolvers/EmbeddingDocTempSolver.java

@@ -0,0 +1,106 @@
+package com.olympus.hermione.stepSolvers;
+
+import ch.qos.logback.classic.Logger;
+import com.olympus.hermione.models.ScenarioExecution;
+import com.olympus.hermione.utility.AttributeParser;
+import java.io.File;
+import java.util.Collections;
+import java.util.List;
+import org.apache.tika.Tika;
+import org.slf4j.LoggerFactory;
+import org.springframework.ai.document.Document;
+import org.springframework.ai.transformer.splitter.TokenTextSplitter;
+
+public class EmbeddingDocTempSolver extends StepSolver {
+    
+    private String scenario_execution_id;
+    private String path_file;
+    private int default_chunk_size;
+    private int min_chunk_size;
+    private int min_chunk_length_to_embed;
+    private int max_num_chunks;
+
+
+
+    Logger logger = (Logger) LoggerFactory.getLogger(BasicQueryRagSolver.class);
+
+    private void loadParameters(){
+        logger.info("Loading parameters");
+        this.scenarioExecution.getExecSharedMap().put("scenario_execution_id", this.scenarioExecution.getId());
+        logger.info("Scenario Execution ID: "+this.scenarioExecution.getId());
+
+        AttributeParser attributeParser = new AttributeParser(this.scenarioExecution);
+
+        this.scenario_execution_id = attributeParser.parse((String) this.scenarioExecution.getId());
+
+        this.path_file = attributeParser.parse((String) this.step.getAttributes().get("path_file"));
+
+        if(this.step.getAttributes().containsKey("default_chunk_size")){
+            this.default_chunk_size = (int) this.step.getAttributes().get("default_chunk_size");
+        }else{
+            this.default_chunk_size = 8000;
+        }
+
+        if(this.step.getAttributes().containsKey("min_chunk_size")){
+            this.min_chunk_size = (int) this.step.getAttributes().get("min_chunk_size");
+        }else{
+            this.min_chunk_size = 50;
+        }
+
+        if(this.step.getAttributes().containsKey("min_chunk_length_to_embed")){
+            this.min_chunk_length_to_embed = (int) this.step.getAttributes().get("min_chunk_length_to_embed");
+        }else{
+            this.min_chunk_length_to_embed = 50;
+        }
+
+        if(this.step.getAttributes().containsKey("max_num_chunks")){
+            this.max_num_chunks = (int) this.step.getAttributes().get("max_num_chunks");
+        }else{
+            this.max_num_chunks = 1000;
+        }
+
+    }
+
+    @Override
+    public ScenarioExecution solveStep(){
+
+        try{
+            logger.info("Solving step: " + this.step.getName());
+            this.scenarioExecution.setCurrentStepId(this.step.getStepId());
+            loadParameters();
+            
+            logger.info("Embedding documents");
+            File file = new File(this.path_file);
+            Tika tika = new Tika();
+            tika.setMaxStringLength(-1);
+            String text = tika.parseToString(file);
+            Document myDoc = new Document(text);
+
+            List<Document> docs = Collections.singletonList(myDoc);
+
+            TokenTextSplitter splitter = new TokenTextSplitter(this.default_chunk_size,
+                    this.min_chunk_size,
+                    this.min_chunk_length_to_embed,
+                    this.max_num_chunks,
+                    true);
+
+            docs.forEach(doc -> {
+                List<Document> splitDocs = splitter.split(doc);
+                logger.info("Number of documents: " + splitDocs.size());
+                
+                splitDocs.forEach(splitDoc -> {
+                    splitDoc.getMetadata().put("KsScenarioExecutionId", this.scenario_execution_id);
+                });
+                vectorStore.add(splitDocs);
+            });
+        }catch (Exception e){
+            logger.error("Error while solvingStep: "+e.getMessage());
+            e.printStackTrace();
+        }
+
+        this.scenarioExecution.setNextStepId(this.step.getNextStepId());
+
+        return this.scenarioExecution;
+    }
+
+}