Merge pull request #157 from rostilos/1.5.3-rc

feat: Add configurable timeout for OpenRouter embedding model

Author: rostilos

deployment/config/rag-pipeline/.env.sample

@@ -50,6 +50,7 @@ REDIS_URL=redis://redis:6379/1
 #OLLAMA_BATCH_SIZE=100
 
 ## Request timeout in seconds (increase for slow CPU)
+#OPENROUTER_TIMEOUT=300
 #OLLAMA_TIMEOUT=120
 
 ## CPU Threading Optimization (set based on your CPU cores)

java-ecosystem/libs/vcs-client/src/main/java/org/rostilos/codecrow/vcsclient/gitlab/GitLabClient.java

@@ -14,6 +14,8 @@
 import java.net.URLEncoder;
 import java.nio.charset.StandardCharsets;
 import java.nio.file.Path;
+import java.time.OffsetDateTime;
+import java.time.format.DateTimeFormatter;
 import java.util.ArrayList;
 import java.util.List;
 
@@ -498,6 +500,78 @@ public String getLatestCommitHash(String workspaceId, String repoIdOrSlug, Strin
         }
     }
 
+    @Override
+    public List<VcsCommit> getCommitHistory(String workspaceId, String repoIdOrSlug, String branchOrCommit, int limit) throws IOException {
+        String projectPath = workspaceId + "/" + repoIdOrSlug;
+        String encodedPath = URLEncoder.encode(projectPath, StandardCharsets.UTF_8);
+        String encodedRef = URLEncoder.encode(branchOrCommit, StandardCharsets.UTF_8);
+        int perPage = Math.min(limit, 100); // GitLab max per_page is 100
+
+        String url = baseUrl + "/projects/" + encodedPath + "/repository/commits?ref_name=" + encodedRef + "&per_page=" + perPage;
+
+        List<VcsCommit> commits = new ArrayList<>();
+
+        while (url != null && commits.size() < limit) {
+            Request request = createGetRequest(url);
+            try (Response response = httpClient.newCall(request).execute()) {
+                if (!response.isSuccessful()) {
+                    throw createException("get commit history", response);
+                }
+
+                JsonNode root = objectMapper.readTree(response.body().string());
+                if (root == null || !root.isArray()) break;
+
+                for (JsonNode commitNode : root) {
+                    if (commits.size() >= limit) break;
+
+                    String hash = getTextOrNull(commitNode, "id");
+                    if (hash == null) continue;
+
+                    String message = getTextOrNull(commitNode, "message");
+                    String authorName = getTextOrNull(commitNode, "author_name");
+                    String authorEmail = getTextOrNull(commitNode, "author_email");
+                    OffsetDateTime timestamp = null;
+
+                    String dateStr = getTextOrNull(commitNode, "authored_date");
+                    if (dateStr != null) {
+                        try {
+                            timestamp = OffsetDateTime.parse(dateStr, DateTimeFormatter.ISO_OFFSET_DATE_TIME);
+                        } catch (Exception e) {
+                            log.debug("Could not parse commit date '{}': {}", dateStr, e.getMessage());
+                        }
+                    }
+
+                    List<String> parentHashes = new ArrayList<>();
+                    JsonNode parentsArray = commitNode.get("parent_ids");
+                    if (parentsArray != null && parentsArray.isArray()) {
+                        for (JsonNode parentId : parentsArray) {
+                            if (!parentId.isNull()) {
+                                parentHashes.add(parentId.asText());
+                            }
+                        }
+                    }
+
+                    commits.add(new VcsCommit(hash, message, authorName, authorEmail, timestamp, parentHashes));
+                }
+
+                // Follow pagination via Link header or X-Next-Page
+                if (commits.size() < limit) {
+                    String nextPage = response.header("X-Next-Page");
+                    if (nextPage != null && !nextPage.isEmpty()) {
+                        url = baseUrl + "/projects/" + encodedPath + "/repository/commits?ref_name=" + encodedRef
+                                + "&per_page=" + perPage + "&page=" + nextPage;
+                    } else {
+                        url = null;
+                    }
+                } else {
+                    url = null;
+                }
+            }
+        }
+
+        return commits;
+    }
+
     @Override
     public String getBranchDiff(String workspaceId, String repoIdOrSlug, String baseBranch, String compareBranch) throws IOException {
         // GitLab: GET /projects/:id/repository/compare

python-ecosystem/rag-pipeline/src/rag_pipeline/core/embedding_factory.py

@@ -4,6 +4,7 @@
 """
 
 import logging
+import os
 from typing import Union
 
 from llama_index.core.base.embeddings.base import BaseEmbedding
@@ -28,31 +29,34 @@ def create_embedding_model(config: RAGConfig) -> BaseEmbedding:
     provider = config.embedding_provider.lower()
 
     if provider == "ollama":
-        logger.info(f"Creating Ollama embedding model: {config.ollama_model}")
+        timeout = float(os.getenv("OLLAMA_TIMEOUT", "120"))
+        logger.info(f"Creating Ollama embedding model: {config.ollama_model} (timeout={timeout}s)")
         return OllamaEmbedding(
             model=config.ollama_model,
             base_url=config.ollama_base_url,
-            timeout=120.0,
+            timeout=timeout,
             expected_dim=config.embedding_dim
         )
 
     elif provider == "openrouter":
-        logger.info(f"Creating OpenRouter embedding model: {config.openrouter_model}")
+        timeout = float(os.getenv("OPENROUTER_TIMEOUT", "300"))
+        logger.info(f"Creating OpenRouter embedding model: {config.openrouter_model} (timeout={timeout}s)")
         return OpenRouterEmbedding(
             api_key=config.openrouter_api_key,
             model=config.openrouter_model,
             api_base=config.openrouter_base_url,
-            timeout=60.0,
+            timeout=timeout,
             max_retries=3,
             expected_dim=config.embedding_dim
         )
 
     else:
         logger.warning(f"Unknown embedding provider '{provider}', defaulting to Ollama")
+        timeout = float(os.getenv("OLLAMA_TIMEOUT", "120"))
         return OllamaEmbedding(
             model=config.ollama_model,
             base_url=config.ollama_base_url,
-            timeout=120.0,
+            timeout=timeout,
             expected_dim=config.embedding_dim
         )
 

python-ecosystem/rag-pipeline/src/rag_pipeline/core/openrouter_embedding.py

@@ -36,7 +36,7 @@ def __init__(
         api_key: str,
         model: str = "openai/text-embedding-3-small",
         api_base: str = "https://openrouter.ai/api/v1",
-        timeout: float = 60.0,
+        timeout: float = float(os.getenv("OPENROUTER_TIMEOUT", "300")),
         max_retries: int = 3,
         embed_batch_size: int = EMBEDDING_BATCH_SIZE,
         expected_dim: Optional[int] = None,