- Introduced a new index `documents_scraped_created_at_id_idx` in the tender repository to support keyset pagination for documents marked as scraped, sorted by creation date.
- This enhancement improves the efficiency of querying scraped documents, facilitating better performance in data retrieval and management.
This update significantly optimizes the tender repository's handling of scraped documents, enhancing overall data access capabilities.
- Introduced a new `translatedNoticesScanner` interface for scanning MinIO for daily translated notice counts.
- Enhanced the `repository` struct to include fields for managing translated notice scope and caching.
- Implemented the `ScanTranslatedNotices` method in the `StorageClient` to retrieve daily counts and total from MinIO.
- Updated the `Statistics` method in the repository to utilize the new translated notices scope, improving data accuracy.
- Added unit tests for the translated notices functionality, ensuring robust validation of the new features.
This update significantly enhances the dashboard's ability to handle translated notice statistics, improving overall data management and reporting capabilities.
- Updated the tender search handler to include a new `include_total` query parameter, allowing clients to request total count and pagination metadata for search results.
- Refactored the `tenderSearchListProjection` to exclude heavy fields from the list response, optimizing data retrieval.
- Modified the `buildSearchFilter` method to utilize the `processing_metadata.documents_scraped` flag for filtering, improving search accuracy.
- Added unit tests for the new pagination features and search filter logic, ensuring robust validation of the search functionality.
This update significantly enhances the tender search experience by providing more flexible pagination options and improving the efficiency of data handling.
- Removed the resetting of `SubmissionMode` in the `Reject` method of the `TenderApproval` entity, ensuring that the submission mode is preserved upon rejection.
- Updated the `GetCompanyTenderApprovalStats` method in the repository to count submissions by mode without filtering by status, enhancing the accuracy of statistics.
- Adjusted the `ToggleTenderApproval` method in the service to set the `SubmissionMode` when rejecting a tender, ensuring consistent handling of submission modes across approval states.
This update improves the clarity and functionality of tender approval processes, ensuring that submission modes are correctly managed during approval and rejection scenarios.
- Introduced `PickAssignedCompanyID` function to determine the appropriate company ID for company-scoped operations, improving company assignment logic.
- Added `ResolveMongoID` method in the tender service to map MongoDB IDs and AI procedure references to canonical tender IDs, enhancing ID resolution.
- Updated `ToggleTenderApproval` to handle company ID validation and improve error handling, ensuring robust processing of tender approvals.
- Enhanced logging for error scenarios in the tender approval service, providing better insights into failures during operations.
- Refactored the `GetByID` and `GetByTenderAndCompany` methods in the tender approval repository to utilize custom error types for improved clarity.
This update significantly improves the handling of company and tender operations, enhancing error management and overall service reliability.
- Introduced a new method `isRecommendedPageCacheStale` in the `tenderService` to determine if the cached recommendations for a company are outdated based on the count of cached recommendations and the latest AI recommendations.
- Updated the `recommendFromPageCache` method to utilize the new cache staleness check, ensuring that outdated caches are not used for recommendations.
- This enhancement improves the accuracy of recommendations by ensuring that users receive the most up-to-date information, thereby enhancing the overall user experience in the tender recommendation service.
- Removed the `buildRecommendedTenderResponsesParallel` method and replaced it with `buildRecommendedTenderListResponses`, simplifying the response building process.
- Introduced `enrichRecommendedTenderResponsesParallel` to handle enrichment of tender responses in parallel, improving performance.
- Updated the `Recommend` method to utilize the new enrichment logic, ensuring timely and efficient processing of recommendations.
- Cleaned up unused functions and imports in `recommendation_page_cache.go` and `recommendation_response.go`, enhancing code clarity and maintainability.
This update streamlines the recommendation response handling and improves the overall efficiency of the tender recommendation service.
- Removed the `RecommendationPageCacheLanguages` configuration from `AISummarizerConfig` to streamline cache management.
- Updated the `companyService` and `tenderService` to utilize the new `InvalidateRecommendedTendersPageCache` method for cache invalidation, enhancing clarity and efficiency.
- Refactored the `invalidateRecommendedTendersPageCache` method to eliminate unnecessary context parameters, simplifying the function signature.
- Improved the handling of page cache refresh logic by consolidating language handling within the `tenderService`, ensuring consistent behavior across services.
- Cleaned up related tests and removed deprecated functions to maintain code quality and readability.
This update enhances the maintainability of the recommendation caching system by simplifying configuration and improving cache invalidation logic.
- Introduced a new `RecommendedTendersPageCacheRefresher` interface to manage the asynchronous refresh of recommendation pages in Redis, improving cache management.
- Updated the `companyService` to support setting page cache languages and refreshing the recommended tenders page cache based on company IDs.
- Enhanced the `tenderService` to build and invalidate recommended tenders page caches, ensuring timely updates and efficient retrieval of cached recommendations.
- Added configuration options for recommendation page cache languages in the `AISummarizerConfig`, allowing for flexible language support.
- Implemented unit tests for the new caching logic and page refresh functionality, ensuring robust validation of the recommendation caching process.
This update significantly improves the efficiency and responsiveness of the tender recommendation service by integrating enhanced caching mechanisms and page refresh capabilities.
- Updated the `markResolvedRecommendedTenderSeen` function to accept an additional `tenderRef` parameter, allowing for better tracking of tender references and preventing duplicates.
- Refactored the `GetByProcedureReferences` method in the repository to combine indexed notice lookups with contract-folder candidates, ensuring older notice references resolve correctly after tender merges.
- Introduced a new `findTendersByContractFolderIDs` method to streamline the retrieval of tenders by contract folder IDs, enhancing performance and maintainability.
- Added unit tests for new functionality in `ai_reference_test.go` and `recommendation_filter_test.go`, ensuring robust validation of tender reference mapping and deduplication logic.
This update significantly improves the handling of tender references and enhances the overall test coverage, ensuring more reliable and efficient processing of tender data.
- Introduced a context with a timeout for Redis cache retrieval in the `getCachedAIRecommendations` method, ensuring more robust handling of potential delays.
- Added a new file `recommendation_page.go` to encapsulate the logic for building recommendation pages, improving code organization and readability.
- Implemented a `buildRecommendationPage` method to handle pagination and batch processing of recommendations, enhancing performance and user experience.
- Created a new test file `recommendation_page_test.go` to validate the functionality of the recommendation page building process, ensuring correct pagination and batch resolution.
This update significantly improves the efficiency and maintainability of the tender recommendation service by optimizing caching and implementing structured pagination handling.
- Renamed the test function to `TestBuildRecommendedTenderListResponsesPreservesOrderWithoutStorage` for clarity and updated its logic to reflect changes in the recommendation response building process.
- Removed unnecessary sleep in the test to enhance performance.
- Refactored the `buildRecommendedTenderResponsesParallel` method to `buildRecommendedTenderListResponses`, simplifying the response building without relying on MinIO translation lookups.
- Updated the repository's `mapProcedureReferencesToTenders` function to improve efficiency by grouping candidates by their contract folder IDs, enhancing the mapping process.
This update streamlines the recommendation response handling and enhances the test suite for better maintainability and performance.
- Replaced the direct fetching of tender details within the `GetTenderApprovalsByCompanyIDWithTender`, `GetTenderApprovalsByStatusWithTender`, and `GetTenderApprovalsBySubmissionModeWithTender` methods with a new approach that loads tender details in bulk using `loadTendersByIDs`.
- Introduced helper functions `fullTenderDetailsFromResponse` and `listTenderDetailsFromResponse` to streamline the mapping of tender responses to `TenderDetails`.
- Added a new file `tender_details.go` to encapsulate tender detail mapping logic, improving code organization and readability.
This update enhances the performance of the tender approval service by reducing redundant calls to fetch tender details, leading to more efficient data handling.
- Introduced a caching mechanism for translation responses in the tender recommendation service, improving efficiency by reducing redundant translation requests.
- Refactored the `enrichWithTranslation` method to support an optional translation cache, allowing for faster retrieval of stored translations.
- Implemented parallel processing for building recommended tender responses, utilizing goroutines to enhance performance and responsiveness.
- Added unit tests to validate the new caching behavior and ensure the correct application of translations in various scenarios.
This update significantly improves the performance and responsiveness of the tender recommendation service by integrating caching and parallel processing, enhancing the overall user experience.
- Updated the company service to include a new method for scheduling the refresh of cached AI recommendations after the AI pipeline execution.
- Introduced a new interface for managing cached recommendation refreshes, improving the separation of concerns within the service layer.
- Enhanced the worker initialization to include Redis client support, allowing for better management of recommendation caching.
- Added functionality to list company IDs with existing recommendation caches, ensuring efficient updates post-pipeline runs.
- Implemented unit tests to validate the new recommendation refresh logic and ensure proper handling of various scenarios.
This update significantly improves the handling of AI recommendations by integrating caching mechanisms with the AI pipeline, enhancing overall system performance and responsiveness.
- Introduced a new function `dedupeRecommendationResponsesByTenderID` to remove duplicate tender recommendations based on their IDs, ensuring unique entries in the recommendation list.
- Updated the `mergeRecommendedTenders` function to utilize the new deduplication logic, streamlining the merging process of recommendation lists.
- Enhanced the `fetchAIRecommendations` method to apply deduplication on the fetched recommendations, improving data integrity.
- Added unit tests for the new deduplication function to validate its behavior and ensure correct handling of various input scenarios.
This update enhances the recommendation handling by ensuring that duplicate tender entries are effectively managed, improving the overall quality of the recommendations provided to users.
- Renamed and refactored the AI pipeline auto run functionality to a daily run, enhancing clarity and purpose.
- Introduced a new `AIPipelineDailyWorker` to manage the daily execution of the AI pipeline, replacing the previous auto run implementation.
- Updated configuration fields and logging messages to reflect the change from auto to daily run, ensuring consistent terminology throughout the codebase.
- Removed the obsolete `ai_pipeline_auto.go` file to streamline the worker structure.
This update improves the maintainability and readability of the AI pipeline management by clearly distinguishing between auto and daily run functionalities.
- Introduced the ability to append external links to company profiles through a new API endpoint.
- Enhanced the `Company` entity to include a `Links` field for storing external resource links.
- Created `AddLinksForm` for validating incoming link data and implemented corresponding logic in the service layer.
- Added error handling for link validation, ensuring only valid URLs are accepted and limiting the number of links to 20.
- Implemented unit tests for link management functions, including sanitization and merging of links.
- Updated relevant API documentation to reflect the new functionality.
This update significantly enhances the company management capabilities by allowing the addition of external links, improving the overall user experience and data richness in company profiles.
- Introduced new functionality in `form_companies.go` to handle company selection from various input formats, including legacy `company_ids` and company summary objects.
- Implemented `UnmarshalJSON` methods for `CreateCustomerForm`, `UpdateCustomerForm`, and `AssignCompaniesForm` to support flexible company data parsing.
- Added unit tests in `form_companies_test.go` to validate the correct unmarshalling of company IDs and summaries, ensuring robust handling of different input scenarios.
- Enhanced the `GetByCompanies` method in the customer repository to directly use string company IDs, improving data retrieval efficiency.
This update significantly improves the handling of company data in customer forms, ensuring accurate processing and validation of company selections, while expanding test coverage for these functionalities.
- Introduced a new test file `recommendation_translation_test.go` to validate the behavior of the tender recommendation service regarding translation handling.
- Implemented tests to ensure that the service correctly applies stored translations and retains original text when no translation is available.
- Enhanced the `Recommend` method in the tender service to utilize a structured approach for managing recommended tenders, improving clarity and maintainability.
This update improves the testing coverage for translation handling in tender recommendations, ensuring accurate responses based on available translations and original content.
- Renamed `buildRejectedTenderIDSetForCompanies` to `buildRejectedTenderIDSetsForCompanies` to better reflect its functionality of returning a map of excluded tender IDs by company.
- Updated the implementation to return a nested map structure for better organization of excluded tender IDs per company.
- Added a new test function `TestIsRecommendedTenderExcludedForAllCompanies` to validate the visibility of tenders based on company-specific exclusions.
- Enhanced existing tests to ensure comprehensive coverage of the new logic for handling rejected tenders.
This update improves the clarity and functionality of the tender recommendation process by refining the handling of rejected tenders and expanding test coverage.
- Introduced a new test function `TestMapProcedureReferencesToTenders` to validate the mapping of procedure references to corresponding tenders.
- Enhanced the `GetByProcedureReferences` method in the tender repository to utilize a more efficient filtering mechanism based on unique contract folder IDs.
- Updated pagination logic to dynamically adjust limits based on the number of unique folders and references, ensuring optimal data retrieval.
This update improves the testing coverage for tender mapping functionality and optimizes the repository's data fetching strategy, enhancing overall performance and reliability.
- Introduced a new method to validate company IDs for AI recommendations, ensuring all requested companies exist before processing.
- Added caching logic to retrieve AI recommendations efficiently, reducing unnecessary calls and improving performance.
- Utilized errgroup for concurrent processing of AI recommendations across multiple companies, enhancing responsiveness.
- Updated the tender service to handle rejected tenders more effectively by incorporating concurrent fetching of rejected tender IDs.
This update significantly improves the AI recommendations process by ensuring accurate company validation and optimizing data retrieval through caching and concurrency, enhancing overall system performance and user experience.
- Added functionality to retrieve merged AI recommendations for multiple companies, improving the relevance of tender suggestions based on company-specific data.
- Introduced normalization functions to clean and deduplicate company IDs, ensuring accurate processing of recommendations.
- Enhanced the company context resolution in customer middleware to support multiple assigned companies, improving the handling of company-specific requests.
- Updated the tender recommendation logic to utilize the new merged recommendations and handle exclusions for rejected tenders accordingly.
- Added unit tests to verify the new recommendation merging logic and company ID normalization, ensuring robust functionality.
This update significantly enhances the tender recommendation process by allowing for more comprehensive and relevant suggestions based on multiple company contexts, improving user experience and satisfaction.
- Updated the tender service to include a new dependency for listing rejected tenders by company, allowing for more refined tender recommendations.
- Introduced a new field in the SearchForm to specify whether to exclude rejected tenders from the recommendation results.
- Enhanced the Recommend method to filter out rejected tenders based on the new exclusion logic, improving the relevance of AI-ranked tender recommendations.
- Added unit tests to verify the exclusion logic for rejected tenders, ensuring robust functionality.
This update improves the tender recommendation process by ensuring that company-rejected tenders are not included in the results, enhancing user experience and satisfaction.
- Enhanced API documentation for `GetTenderApprovalByTenderAndCompany` to clarify the default behavior of the status filter, indicating it defaults to 'submitted' when omitted.
- Updated the `PublicGetTenderApprovals` method to set the default status to 'submitted' if no status is provided, ensuring consistent behavior in approval retrieval.
This update improves the clarity of the API documentation and enhances the functionality of the tender approval retrieval process by establishing a default status, thereby improving user experience.
- Added new field `Days` and `ScrapedTED` to `SummaryResponse` for tracking daily TED scrape counts.
- Updated `SummaryQuery` to include `Days` parameter for querying scraped TED data.
- Modified `Summary` handler to accept and process the new `Days` parameter.
- Refactored `Summary` method in the repository to support the new `Days` parameter and improved aggregation logic.
- Enhanced caching logic in the service layer to utilize a composite cache key based on `closingWindowSec` and `days`, improving cache management and retrieval efficiency.
This update improves the dashboard's functionality by providing more detailed insights into TED scraping activities and optimizing the caching strategy for better performance.
- Introduced InvalidStatusTransitionError to handle invalid status transitions for inquiries, providing clearer error messages.
- Updated UpdateInquiryStatusForm to make the Reason field optional and added logic to set a default reason if not provided.
- Enhanced form validation tests to cover new status transition error scenarios and validation messages.
- Refactored handler methods to utilize new error handling functions for improved response management.
This update improves the robustness of inquiry status management by ensuring proper error handling and validation, enhancing user experience during status updates.
- Modified CompanyForm fields (RegistrationNumber, TaxID, Industry) to be optional, allowing for more flexible company creation.
- Updated the Create method in company service to check for existing companies by registration number and tax ID only if provided, improving error handling and user experience during company creation.
This update enhances the company creation process by allowing optional fields while ensuring that existing company checks are performed only when necessary.
- Introduced AIPipelineAutoWorker to manage the execution of the AI pipeline auto run, including startup catch-up and scheduled tasks.
- Enhanced WorkerConfig to include AIPipelineAutoEnabled and AIPipelineAutoInterval settings for better control over AI pipeline execution.
- Added logging for AI pipeline auto run status, including success and error handling, to improve observability.
- Updated daily job tracker to include AIPipelineAutoJobName for tracking AI pipeline job completions.
This update enhances the system's capability to automate AI pipeline executions, improving efficiency and reliability in processing AI tasks.
- Updated the ListPendingTenders method to filter tenders based on their active publication submission window instead of just deadlines, enhancing the accuracy of tender listings.
- Introduced a new HasActivePublicationWindow method in the Tender entity to determine if the publication-based submission window is still open.
- Modified the GetTenderByNoticeID method to reflect the new logic for filtering tenders based on their publication status.
- Enhanced the documentation for relevant API endpoints to clarify the changes in tender retrieval criteria.
This update improves the precision of tender scraping by ensuring only relevant tenders with active submission windows are processed, contributing to better data management and scraping efficiency.
- Added a new field, ScrapedTEDNotices, to the StatisticsLifetimeTotals struct to track the total number of TED notices scraped.
- Updated the Statistics method in the statistics repository to include a background process for retrieving total scraped TED notices, improving the accuracy of dashboard statistics.
- Introduced new methods in the Counter to increment and retrieve daily counts for scraped TED notices, ensuring reliable metrics for reporting.
- Modified the TEDScraper to increment the TED notice scraped counter upon successful import, enhancing the tracking of scraping activity.
This update improves the dashboard's statistics by providing detailed insights into TED notice scraping activities, contributing to better data visibility and reporting.
- Introduced CompanyContextMiddleware to resolve the active company context for customer requests, ensuring that tender recommendations and company-scoped APIs remain in sync with the database.
- Updated public routes to utilize the new CompanyContextMiddleware alongside the existing AuthMiddleware, improving the handling of company-specific requests.
- Added unit tests for the pickActiveCompanyID function to validate the logic for selecting the appropriate company context based on customer assignments and requested company IDs.
This update enhances the accuracy and reliability of company context management in the application, improving user experience and data consistency.
- Introduced a new constant, statisticsColdLoadWait, to define the wait time for cold-cache requests before serving a placeholder response.
- Updated the Statistics method to initiate a background load for statistics while allowing for a brief wait for real data, improving responsiveness during cache warming.
- Implemented a channel to handle the result of the background loading process, ensuring that users receive timely feedback while the cache is being populated.
This update optimizes the dashboard's performance by ensuring that users are served quickly, even when the statistics are being freshly loaded from the backend.
- Updated the Statistics method in the dashboard service to initiate a background refresh for statistics while serving a placeholder report, enhancing responsiveness during cache warming.
- Removed blocking calls to load statistics directly, allowing for faster response times.
- Improved logging to indicate when placeholder statistics are served, providing better visibility into the caching process.
This update optimizes the dashboard's performance by ensuring that users receive immediate feedback while the system prepares accurate statistics in the background.
- Updated the dashboard service to integrate Redis caching for improved performance in statistics retrieval.
- Modified the NewService function to accept a Redis client, enabling caching of dashboard statistics.
- Implemented logic to retrieve statistics from Redis, falling back to the database if necessary, and introduced a background process to warm the cache.
- Enhanced error handling and logging for Redis operations to ensure robust statistics management.
- Increased cache duration for scraped documents and adjusted timeout settings for MongoDB queries to optimize performance.
This update significantly improves the responsiveness and efficiency of the dashboard by leveraging Redis for caching statistics.
- Increased the cache duration for dashboard statistics from 60 seconds to 5 minutes, improving data freshness and reducing load on the backend.
- Introduced a stale cache mechanism that allows retrieval of stale statistics while refreshing them in the background, enhancing user experience by providing quicker access to data.
- Updated the statistics repository to handle the new caching logic, ensuring accurate and timely statistics reporting.
- Added tests to validate the new caching behavior and ensure the integrity of statistics retrieval.
This update optimizes the dashboard's performance and responsiveness by improving the caching strategy for statistics.
- Updated the ListPaginationOptions struct to include SkipCount and IncludeCount fields, allowing for more flexible pagination behavior.
- Modified the BuildListPagination function to handle cursor pagination with count options, improving performance by running count queries in parallel with data retrieval.
- Enhanced the FindAll method in the repository to support concurrent counting of documents, reducing overall latency for list operations.
- Added tests for pagination behavior, ensuring accurate handling of count scenarios in both offset and cursor pagination.
This update improves the efficiency and flexibility of pagination in the MongoDB repository, enhancing the overall performance of list operations.
- Added the PublicationSubmissionDeadline method to the Tender entity, which calculates the submission deadline based on the publication date and stored submission deadline.
- Updated the IsRecommendable method to utilize the new PublicationSubmissionDeadline logic for determining recommendation eligibility.
- Introduced unit tests for the PublicationSubmissionDeadline method, ensuring accurate calculation and validation of submission deadlines.
- Enhanced existing tests for the IsRecommendable method to cover new scenarios related to submission deadlines.
This update improves the accuracy of submission deadline handling and enhances the recommendation logic for tenders based on publication dates.
- Implemented the IsRecommendable method in the Tender entity to determine if a tender should be included in AI recommendations based on its status and deadline.
- Added unit tests for the IsRecommendable method to cover various scenarios, ensuring accurate recommendation logic.
- Updated the Recommend method in the tender service to utilize the new IsRecommendable method for improved clarity and functionality.
This update enhances the recommendation logic for tenders, ensuring only appropriate tenders are considered for AI recommendations based on their status and deadlines.
- Updated error handling in the Elasticsearch client to utilize a new `readErrorBody` function for better diagnostics when pinging and searching.
- Refactored the `flushBatch` method to marshal bulk index actions using a dedicated `marshalBulkIndexAction` function, improving code clarity and error handling.
- Enhanced the overall structure of the Elasticsearch client for improved maintainability and readability.
This update enhances the robustness of the Elasticsearch client, ensuring more informative error messages and cleaner code organization.
- Introduced a new `auditlog` package to handle audit logging for user actions, including creation, updates, deletions, and authentication events.
- Enhanced existing services (customer, user) to log relevant actions using the new audit logger, capturing details such as actor ID, action type, target type, and success status.
- Added middleware to enrich request context with metadata for audit logging, ensuring comprehensive tracking of user interactions.
- Integrated Elasticsearch for persistent storage of audit logs, with fallback to file-only logging if Elasticsearch is unavailable.
- Updated API documentation to include new audit log endpoints for administrative access.
This update significantly improves the system's ability to track and audit user actions, enhancing security and accountability within the application.
- Added a new `UserID` field to the `SearchForm` to allow filtering notifications by user ID.
- Implemented the `ResolvedRecipients` method to return user IDs based on the `user_id` or `recipient` query parameters, improving the flexibility of notification searches.
- Updated the `GetNotifications` method in the service layer to utilize the new recipient resolution logic, ensuring accurate retrieval of notifications based on user ID.
This update enhances the notification management capabilities, providing more granular control over notification searches.
- Introduced a new `.gitattributes` file to ensure consistent handling of text files across different platforms by setting the `text=auto` attribute.
- This addition helps maintain cross-platform compatibility and prevents potential issues with line endings in text files.
This update lays the groundwork for better version control practices in the project, enhancing collaboration among developers working on different operating systems.
- Introduced the `scrapedDocumentsScanner` interface to facilitate scanning of scraped documents from MinIO, returning both procedure summaries and daily document counts.
- Updated the `ListProceduresWithDocuments` method to utilize the new scanning functionality, improving data retrieval efficiency.
- Enhanced the `scrapedDocumentsPerDay` method to filter daily counts based on a specified start date, ensuring accurate reporting of document statistics.
- Added unit tests for the new scanning logic and daily counts filtering, ensuring robust functionality and error handling.
This update enhances the dashboard's document management capabilities, providing better insights into scraped documents and their daily counts.
- Introduced a mutex to ensure only one TED scraper run executes at a time, preventing concurrent executions during startup and scheduled runs.
- Implemented a mechanism to check if today's TED scrape has already been completed during startup, logging appropriate messages for both completed and new runs.
- Added startup catch-up logic for tender translations and unprocessed notices, ensuring that any missed tasks are executed without blocking the application startup.
This update improves the reliability and efficiency of the TED scraper and worker processes, ensuring that all necessary tasks are completed after a server restart.
- Introduced retry logic for fetching AI recommendations after onboarding, enhancing reliability in recommendation retrieval.
- Updated logging levels for better observability, changing cache miss logs to Info level.
- Renamed methods for clarity, replacing `refreshAIRecommendationsCacheAsync` with `scheduleRecommendationRefreshAfterOnboarding` and `fetchAndCacheAIRecommendations` with `fetchAIRecommendations`.
- Implemented a mechanism to clear the cache if no recommendations are returned, improving cache management.
This update optimizes the AI recommendation process, ensuring more robust handling of recommendation fetching and caching during onboarding.
- Updated `AISummarizerConfig` to allow for a default `RecommendationCacheTTL` of 0, enabling persistent caching until company updates.
- Refactored `StartAIOnboarding` to include cache invalidation and asynchronous recommendation refresh, improving responsiveness during onboarding.
- Introduced `triggerAIOnboardingAsync` method for background processing of AI onboarding and cache refresh, enhancing user experience.
- Improved logging for AI onboarding and recommendation fetching processes, providing better observability and error tracking.
This update optimizes the AI recommendation caching mechanism and onboarding workflow, ensuring a smoother and more efficient experience for users.
Mazyar