# Overview of the Job recommendation system, connection between entities in the _Forcepush_ domain

skinparam linetype polyline
skinparam maxMessageSize 100
'skinparam SequenceMessageAlign center

title Job recommendation system overview


Actor jobseeker

rectangle users {
    Cloud site
    Cloud email
}

jobseeker --> site
email --> jobseeker


cloud "Internet" {

    rectangle ExternalAPIs
    rectangle PartnerFeeds

    rectangle "Old partner feeds\n not visible to preprocess" as oldpartnerfeeds
    package S3 {
        rectangle "Organic feed" as organic_feed
    }
}



cloud AWS {
    database Dynamo as dynamo <<dynamodb>>
    database "Reporting" as reporting <<mysql>>

}

package "Forcepush domain" as fp {
   package Jobrec as JobrecPackage {
        () jobrec <<http>>
        
        database "Personalization db" as jrec_personalize_cache <<redis>>
        database "Dedup cache" as jrec_dedup_cache <<redis>>
    }

    package Match {
        package "Compare " {
            () " " as compare <<http>>
            [Title vectorizer]
        }
        package "Recommend" {
            () " " as recommend <<http>> 
        }

        [preprocess] as preprocess
        note "Multiple feed transformation tasks\n running in parallel, \nprocessing partner feeds and \norganic feed as well. \nIn the meantime enriches job infromation." as preprocess_note
        preprocess ..> preprocess_note
        recommend -[hidden]-> preprocess
        database JobDb <<ElasticSearch>>

    }




    rectangle "Job Enrichment related services" {

        package "Job Enricher" as je {
            database "Job Enricher Cache" as jecache <<redis>>
            () " " as jehttp <<http>>
        }

        package "Job Classifier" as jc {
            () " " as jchttp <<http>>
        }

        package "Company Enricher" as ce {
            () " " as cehttp <<http>>
            database companydb <<redis>>
        }
        
        package "Magellan Geocoder" as magellan {
            () " " as maghttp <<http>>
        }
        

     }

    rectangle "Crawlers" {

        database crawlerDb <<ElasticSearch>>

        package "Job Crawler Front End" as jcfe {

        }

        package "Job Crawler Back End" as jcbe {

        }

        jcfe -> crawlerDb : read
        jcbe --> crawlerDb : write

        jcfe -> organic_feed  : write this feed as xml
        jcbe -> oldpartnerfeeds : process these feeds

        jcfe -> dynamo  : write the jobs properties that are included in the sitemap/xml
    }

}

preprocess -( jehttp  : enriches each job
preprocess ---( maghttp : normalize location info
preprocess -> organic_feed  : reads
preprocess -> PartnerFeeds : reads

jobrec --( recommend  : get jobs
jobrec --( compare  : vectorize 3rd party apis
jobrec -> ExternalAPIs

'JobrecPackage -> reporting


je --( jchttp 
je --( cehttp 




recommend --> JobDb  : read
preprocess -up-> JobDb  : write



site --( jobrec
email --( jobrec

site .> dynamo  : Check the job in the job details page (== sitemap) url. Returns job properties,these will be somewhat mapped to the jobrec request.

## Crawlers
### Job Crawler Backend
Async
### Job Crawler Frontend
Async

## Jobrec
Sync

## Match
### Recommend
Sync
### Compare
Sync
### Preprocess
Async

## Job Enrichment related services
### Job Enricher
Sync

#### Company Enricher
Sync
#### Title classifier
Sync