High level processes

[mjuchli]

Dear all,

According to the lecture slides, we see two processes of an IR system:

• Indexing
• Query and result presentation

Combining this with my thoughts about structuring the project, I came up with a separation of three processes for now:

• Content Acquisition
• Content Processing
• Search process

Those three steps build a high-level pipeline of an IR system. If we zoom into each of those components, we should be able to draw an internal pipeline. A rudimentary draft of such an internal pipeline was already created for the Search process.

Further more, I believe that the segmentation of the content acquisition and processing will help us to better separate our tasks and the therefore related concerns. Likewise I could imagine that we can split up the search process in the parts:

• User Interface / Query Processing
• Search / Ranking

[mjuchli]

Initial draft provided with: 264f9c19aad0e1388ed78c3f6680d3b3e1bd7931 and 310d5156d9d528df96c8a552fa19f3e68d766c04

[lexrazoux]

20170302LRS_Pipeline.pptx

[mjuchli]

@Krymnos and me created the issues for the project idea document. Among other things, that contains indexing process (#16) and Query and result presentation process (#17). The first of which is a merge of previously created Content Acquisition as well as Content Processing. The latter one is basically the same as Search Processing.