Part 1: CRUD
Nothing important, simple CRUD.
ElasticSearch is basically Apache Lucene with scalability. It provids scalability as in MongoDB: Read/Write core Nodes as Masters, Read-only Nodes as Slaves. Information can be parsed in parallel, data can be sharded by nodes and clusters.
Elastic Stack is Stack of 4 components:
- ElasticSearch: search engine
- Logstash: a light-weight, open-source, server-side data processing pipeline that allows you to collect data from a variety of sources, transform it on the fly, and send it to your desired destination.
- Kibana: Frontend for visualization, management and analytics
- Beats: Single-purpose data shipper for ES.
Part 2: Relevance of a search
How do we measure relevance?
- Precision - favors quality over quantity
- Recall - favors quantity over quality
What is relevance of a search?
- True positives - relevant documents that are returned to the user
- False positives - NOT relevant documents that are returned to the user
- True negatives - NOT relevant document that are NOT returned to the user
- False negatives - relevant document that are NOT returned to the user
What is ranking?
Every doc has a score, higher is better
What is score?
Score - value that represents how relevant a document is to that specific query
Score is computed for each document that is a hit
There are multiple types of analysis, but there it will be focused only on two:
Term Frequency
Inverse Document Frequency
Key terms
- minimum should match
- AND/OR operators
- Aggregation (very strong)
Part 3: Full text queries
Match queries
- match query - basic one
- match_phrase query - matches phrases
- multi_match - match only in specific fields and score increase (^2), can add phrase type match
- bool query
Part 4: Aggregations
- metric: lowest unit, highest unit, filtration, average, stats, sum or all at the same time
- cardinality: number of unique values.
Bucket aggregation
- Date Histogram: fixed_interval: “2015-08-14”, calendar_interval: “1M”
- Order aggregation: descending, ascending, so on
- Histogram aggregation
- Transactions per price interval
- Range aggregation
- Terms aggregation: count on unique terms?
Part 5: Mapping
What is mapping?
it’s like elasticsearch db schema. Names and types of an index
Dynamic mapping
when a user does not define it manually, ES does it automatically.
Types:
- Text - for full-text searches
- Keyword - exact searches, aggregations, sorting (like categories)
Runtime field
You can actually create a temporary field to run any request that was done on runtime (IDK)