Looking for a lightweight, yet scalable solution for implementing a full text search index in JavaScript using static files accessible via HTTP? Seeking to make about 100k documents searchable online without breaking the bank on hosting costs like Elasticsearch or private Google search servers? With limited resources, but the ability to host JSON and simple text files inexpensively, I'm exploring ways to create a basic search engine. Preferably, I'd like something that caters to simple keyword searches without complex query languages.
One approach I've considered involves parsing all documents, creating bag-of-words representations for each file, and generating index files listing document IDs and word counts. For search functionality, a straightforward JavaScript or Python script would retrieve index files for user queries, identify document IDs with the highest term counts, and generate search results accordingly.
While cost-effective and feasible for my needs, this method has its limitations in terms of efficiency due to the size and processing requirements of index files. Despite researching extensively, I haven't come across similar client-side solutions utilizing server-generated static index files. Existing options either involve expensive full text search servers or loading large indexes on the client side, neither of which are viable given my constraints.
I'm open to suggestions on optimizing the structure of index files or discovering more efficient tools or approaches for this type of search implementation. Any insights or recommendations would be greatly appreciated!