718 create a single reactor to search both engines and apps

[shubhammahure]

MyEngineProjectReactor

Description

The MyEngineProjectReactor is a unified reactor that consolidates the logic for fetching data from both Engines and Projects. It determines the source of data based on the type parameter and executes a search accordingly.
Support for new metadata fields TYPE and SUB_TYPE has been added to the description section. These fields help filter and classify engines and projects more accurately during the search.

Conditional Fetching

If type is provided in the parameters, it performs a search for Engines.

If type is not provided, it fall to else part for searching with Projects.

How to Test(via app terminal)

For Engine

MyEngineProject( metaKeys = ["tag","domain","data classification","data restrictions","description"] , metaFilters = [ {} ] , filterWord=[""], type=['ENGINE'], sub_type=['MODEL'], offset=[0], limit=[15]) ;

For Project(App)

MyEngineProject(metaKeys = ["tag","domain","data classification","embedder","data restrictions","description"], metaFilters=[{}], filterWord=[""], type=['APP'], sub_type=[{}], onlyPortals=[true]);

A screenshot has been added for better clarity.

[github-actions]

@CodiumAI-Agent /describe

[QodoAI-Agent]

Title

718 create a single reactor to search both engines and apps

---

User description

MyEngineProjectReactor

Description

The MyEngineProjectReactor is a unified reactor that consolidates the logic for fetching data from both Engines and Projects. It determines the source of data based on the type parameter and executes a search accordingly.

Conditional Fetching

If engineType is provided in the parameters, it performs a search for Engines.

If engineType is not provided, it fall to else part for searching with Projects.

How to Test(via app terminal)

For Engine

MyEngineProject( metaKeys = ["tag","domain","data classification","data restrictions","description"] , metaFilters = [ {} ] , filterWord=[""], userT = [true], engineTypes=['MODEL'], offset=[0], limit=[15]) ;

For Project(App)

MyEngineProject(metaKeys = ["tag","domain","data classification","data restrictions","description"], metaFilters=[{}], filterWord=[""], onlyPortals=[true]);

A screenshot has been added for better clarity.

---

PR Type

Enhancement

---

Description

Add unified engine/project search reactor
Implement conditional fetching based on engineType
Enrich results with metadata and user tracking
Provide helper methods for filter extraction

---

Changes walkthrough 📝

	Relevant files
Enhancement	MyEngineProjectReactor.java Unified engine and project search reactor                                src/prerna/reactor/MyEngineProjectReactor.java Added new reactor for unified engine/project search Implemented conditional logic based on engineType presence Enriched results with metadata and user tracking Added helper methods extracting filter parameters +319/-0

---

Need help?

Type /help how to ... in the comments thread for any questions about PR-Agent usage.

Check out the documentation for more information.

[github-actions]

@CodiumAI-Agent /review

[github-actions]

@CodiumAI-Agent /improve

[QodoAI-Agent]

PR Reviewer Guide 🔍

Here are some key observations to aid the review process:

⏱️ Estimated effort to review: 4 🔵🔵🔵🔵⚪

🧪 No relevant tests

🔒 No security concerns identified

⚡ Recommended focus areas for review Null Safety Filter and metadata helper methods return null instead of empty collections, which can lead to NullPointerExceptions downstream when these are dereferenced. private List<String> getEngineIdFilters() { GenRowStruct grs = this.store.getNoun(ReactorKeysEnum.ENGINE.getKey()); if(grs != null && !grs.isEmpty()) { return grs.getAllStrValues(); } return null; } Code Duplication The loops for fetching and merging metadata in both the engine and project branches are nearly identical; consider extracting them into a shared helper to reduce duplication and simplify maintenance. if(!noMeta) { IRawSelectWrapper wrapper = null; try { wrapper = SecurityEngineUtils.getEngineMetadataWrapper(index.keySet(), getMetaKeys(), true); while(wrapper.hasNext()) { Object[] data = wrapper.next().getValues(); String databaseId = (String) data[0]; String metaKey = (String) data[1]; String metaValue = (String) data[2]; if(metaValue == null) { continue; } int indexToFind = index.get(databaseId); Map<String, Object> res = engineInfo.get(indexToFind); // whatever it is, if it is single send a single value, if it is multi send as array if(res.containsKey(metaKey)) { Object obj = res.get(metaKey); if(obj instanceof List) { ((List) obj).add(metaValue); } else { List<Object> newList = new ArrayList<>(); newList.add(obj); newList.add(metaValue); res.put(metaKey, newList); } } else { res.put(metaKey, metaValue); } } } catch (Exception e) { classLogger.error(Constants.STACKTRACE, e); } finally { if(wrapper!=null) { try { wrapper.close(); } catch (IOException e) { classLogger.error(Constants.STACKTRACE, e); } } } Missing Feature The includeUserT flag is only applied for engines; the project branch does not fetch user tracking or vote data, leading to inconsistent behavior between the two flows. if(!projectInfo.isEmpty() && (!noMeta || includeUserT)) { Map<String, Integer> index = new HashMap<>(projectInfo.size()); int size = projectInfo.size(); // now we want to add most executed insights for(int i = 0; i < size; i++) { Map<String, Object> project = projectInfo.get(i); String projectId = project.get("project_id").toString(); // keep list of project ids to get the index index.put(projectId, Integer.valueOf(i)); } if(!noMeta) { IRawSelectWrapper wrapper = null; try { wrapper = SecurityProjectUtils.getProjectMetadataWrapper(index.keySet(), getMetaKeys(), true); while(wrapper.hasNext()) { Object[] data = wrapper.next().getValues(); String projectId = (String) data[0]; String metaKey = (String) data[1]; String metaValue = (String) data[2]; if(metaValue == null) { continue; } int indexToFind = index.get(projectId); Map<String, Object> res = projectInfo.get(indexToFind); // whatever it is, if it is single send a single value, if it is multi send as array if(res.containsKey(metaKey)) { Object obj = res.get(metaKey); if(obj instanceof List) { ((List) obj).add(metaValue); } else { List<Object> newList = new ArrayList<>(); newList.add(obj); newList.add(metaValue); res.put(metaKey, newList); } } else { res.put(metaKey, metaValue); } } } catch (Exception e) { classLogger.error(Constants.STACKTRACE, e); } finally { if(wrapper != null) { try { wrapper.close(); } catch (IOException e) { classLogger.error(Constants.STACKTRACE, e); } } } } } return new NounMetadata(projectInfo, PixelDataType.CUSTOM_DATA_STRUCTURE, PixelOperationType.PROJECT_INFO);

[QodoAI-Agent]

PR Code Suggestions ✨

Latest suggestions up to 053ee8c

Category	Suggestion	Impact
Possible issue	Return empty list instead of null Returning null from collection getters can cause NullPointerExceptions in downstream code; return an empty list instead. src/prerna/reactor/MyEngineProjectReactor.java [227-234] private List<String> getEngineIdFilters() { GenRowStruct grs = this.store.getNoun(ReactorKeysEnum.ENGINE.getKey()); - if(grs != null && !grs.isEmpty()) { + if (grs != null && !grs.isEmpty()) { return grs.getAllStrValues(); } - - return null; + return Collections.emptyList(); } Suggestion importance[1-10]: 6 __ Why: Returning null from getEngineIdFilters can lead to NPEs downstream; returning Collections.emptyList() is a safe, minimal-impact improvement.	Low
	Check map lookup result before use Guard against missing keys in the index map to avoid NullPointerExceptions when get returns null; skip entries that don't exist in the map. src/prerna/reactor/MyEngineProjectReactor.java [78-86] String databaseId = (String) data[0]; -int indexToFind = index.get(databaseId); +Integer indexToFind = index.get(databaseId); +if (indexToFind == null) { + continue; +} Suggestion importance[1-10]: 6 __ Why: Guarding the index.get(databaseId) call avoids potential NPEs if a key is missing, adding a simple, defensive check.	Low
General	Use try-with-resources for wrapper Use a try-with-resources block for IRawSelectWrapper to simplify resource management and ensure close() is called automatically. src/prerna/reactor/MyEngineProjectReactor.java [73-113] -IRawSelectWrapper wrapper = null; -try { - wrapper = SecurityEngineUtils.getEngineMetadataWrapper(index.keySet(), getMetaKeys(), true); - while(wrapper.hasNext()) { +try (IRawSelectWrapper wrapper = + SecurityEngineUtils.getEngineMetadataWrapper(index.keySet(), getMetaKeys(), true)) { + while (wrapper.hasNext()) { ... } } catch (Exception e) { classLogger.error(Constants.STACKTRACE, e); -} finally { - if(wrapper!=null) { - try { - wrapper.close(); - } catch (IOException e) { - classLogger.error(Constants.STACKTRACE, e); - } - } } Suggestion importance[1-10]: 6 __ Why: Converting the manual try/finally close pattern to a try-with-resources block simplifies code and ensures wrapper.close() is always called.	Low

---

Previous suggestions

Suggestions up to commit 053ee8c

Category	Suggestion	Impact
General	Add user tracking for projects You currently only handle metadata for projects but skip user tracking. Add a block similar to the engine branch under includeUserT to fetch and populate upvotes and hasUpvoted flags for projects when tracking is enabled. src/prerna/reactor/MyEngineProjectReactor.java [158-216] List<Map<String, Object>> projectInfo = SecurityProjectUtils.getUserProjectList( this.insight.getUser(), projectIdFilters, favoritesOnly, portalsOnly, engineProjectMetadataFilter, permissionFilters, searchTerm, limit, offset); if(!projectInfo.isEmpty() && (!noMeta || includeUserT)) { Map<String, Integer> index = new HashMap<>(projectInfo.size()); - // metadata block... + // build index... if(!noMeta) { + // existing metadata logic... + } + if(includeUserT && Utility.isUserTrackingEnabled()) { IRawSelectWrapper wrapper = null; try { - wrapper = SecurityProjectUtils.getProjectMetadataWrapper( - index.keySet(), getMetaKeys(), true); - // populate metadata... + wrapper = UserCatalogVoteUtils.getAllVotesWrapper(index.keySet()); + while(wrapper.hasNext()) { + Object[] data = wrapper.next().getValues(); + String projectId = (String) data[0]; + int upvotes = ((Number) data[1]).intValue(); + projectInfo.get(index.get(projectId)).put("upvotes", upvotes); + } } finally { if(wrapper != null) { wrapper.close(); } + } + Map<String, Boolean> voted = UserCatalogVoteUtils.userProjectVotes( + User.getUserIdAndType(this.insight.getUser()), index.keySet()); + for (String pid : index.keySet()) { + projectInfo.get(index.get(pid)).put("hasUpvoted", + voted.getOrDefault(pid, false)); } } } return new NounMetadata(projectInfo, PixelDataType.CUSTOM_DATA_STRUCTURE, PixelOperationType.PROJECT_INFO); Suggestion importance[1-10]: 8 __ Why: The project branch omits the user tracking logic present in the engine branch, so adding upvotes and hasUpvoted maintains feature parity and correct functionality.	Medium
Possible issue	Return empty list instead of null Avoid returning null from filter helpers to prevent NullPointerExceptions when downstream code calls methods like keySet() or iterates over the results. Return an empty list instead of null so callers can safely iterate without additional null checks. src/prerna/reactor/MyEngineProjectReactor.java [227-234] private List<String> getEngineIdFilters() { GenRowStruct grs = this.store.getNoun(ReactorKeysEnum.ENGINE.getKey()); - if(grs != null && !grs.isEmpty()) { + if (grs != null && !grs.isEmpty()) { return grs.getAllStrValues(); } - - return null; + return Collections.emptyList(); } Suggestion importance[1-10]: 6 __ Why: Using Collections.emptyList() prevents potential NPEs when callers iterate the filter results and is a safe default.	Low
	Default metadata map to empty If getMetaMap() returns null, passing it to utility methods may cause NPEs. Default the metadata map to an empty HashMap when null so downstream calls can operate without additional checks. src/prerna/reactor/MyEngineProjectReactor.java [52] Map<String, Object> engineProjectMetadataFilter = getMetaMap(); +if (engineProjectMetadataFilter == null) { + engineProjectMetadataFilter = new HashMap<>(); +} Suggestion importance[1-10]: 5 __ Why: Guarding against null for engineProjectMetadataFilter helps avoid downstream NPEs with minimal impact.	Low

[ppatel9703]

@kunal0137 @ryanweiler92 This works as described. However have a few questions

1. Do we want this to combine the search into one reactor
2. Having different inputs be valid as part of the ask concerns me
3. Do we want user tracking to be passed in as a boolean? Seems like we should just base that off if its enabled in the rdf map

@shubhammahure might have you refactor a few portions of the file based on feedback from above and other small changes

[ppatel9703]

@shubhammahure can you update the description to work for the changes you pushed up

[punithkv4334]

@kunal0137 @ryanweiler92 This works as described. However have a few questions

1. Do we want this to combine the search into one reactor
2. Having different inputs be valid as part of the ask concerns me
3. Do we want user tracking to be passed in as a boolean? Seems like we should just base that off if its enabled in the rdf map

@shubhammahure might have you refactor a few portions of the file based on feedback from above and other small changes

Hi @ppatel9703 ,
I've updated the PR with the changes based on the above feedback.

[github-actions]

@CodiumAI-Agent /update_changelog

[QodoAI-Agent]

Changelog updates: 🔄

2025-05-19 *

Added

• Unified reactor to search both engines and apps
• Support for TYPE and SUB_TYPE metadata filtering

to commit the new content to the CHANGELOG.md file, please type:
'/update_changelog --pr_update_changelog.push_changelog_changes=true'