Skip to content
main
Switch branches/tags
Code

Latest commit

* support prometheus

* format code

* support prometheus
41bad30

Git stats

Files

Permalink
Failed to load latest commit information.

Note: Zinc and all its APIs are considered to be alpha stage at this time. Expect breaking changes in API contracts and data format before at this stage.

Zinc Search Engine

Zinc is a search engine that does full text indexing. It is a lightweight alternative to Elasticsearch and runs using a fraction of the resources. It uses bluge as the underlying indexing library.

It is very simple and easy to operate as opposed to Elasticsearch which requires a couple dozen knobs to understand and tune.

It is a drop-in replacement for Elasticsearch if you are just ingesting data using APIs and searching using kibana (Kibana is not supported with zinc. Zinc provides its own UI).

Check the below video for a quick demo of Zinc.

Zinc Youtube

Join slack channel

Slack

Why zinc

While Elasticsearch is a very good product, it is complex and requires lots of resources and is more than a decade old. I built Zinc so it becomes easier for folks to use full text search indexing without doing a lot of work.

Features:

  1. Provides full text indexing capability
  2. Single binary for installation and running. Binaries available under releases for multiple platforms.
  3. Web UI for querying data written in Vue
  4. Compatibility with Elasticsearch APIs for ingestion of data (single record and bulk API)
  5. Out of the box authentication
  6. Schema less - No need to define schema upfront and different documents in the same index can have different fields.
  7. Index storage in s3 (experimental)
  8. aggregation support

Roadmap items:

  1. High Availability
  2. Distributed reads and writes
  3. Geosptial search
  4. Raise an issue if you are looking for something.

Screenshots

Search screen

Search screen 1 Search screen for games

User management screen

Users screen

Getting started

Download / Installation / Run

Windows

Binaries can be downloaded from releases page for appropriate platform.

C:\> set FIRST_ADMIN_USER=admin
C:\> set FIRST_ADMIN_PASSWORD=Complexpass#123
C:\> mkdir data
C:\> zinc.exe

MacOS - Homebrew

$ brew tap prabhatsharma/tap

$ brew install prabhatsharma/tap/zinc

$ mkdir data

$ FIRST_ADMIN_USER=admin FIRST_ADMIN_PASSWORD=Complexpass#123 zinc

Now point your browser to http://localhost:4080 and login

MacOS/Linux Binaries

Binaries can be downloaded from releases page for appropriate platform.

Create a data folder that will store the data

$ mkdir data

$ FIRST_ADMIN_USER=admin FIRST_ADMIN_PASSWORD=Complexpass#123 ./zinc

Now point your browser to http://localhost:4080 and login

Docker

Docker images are available at https://gallery.ecr.aws/prabhat/zinc

$ mkdir data

$ docker run -v /full/path/of/data:/data -e DATA_PATH="/data" -p 4080:4080 -e FIRST_ADMIN_USER=admin -e FIRST_ADMIN_PASSWORD=Complexpass#123 --name zinc public.ecr.aws/prabhat/zinc:latest

Now point your browser to http://localhost:4080 and login

Kubernetes

Manual Install

Create the namespace:

$ kubectl create ns zinc

$ kubectl apply -f k8s/kube-deployment.yaml

$ kubectl -n zinc port-forward svc/z 4080:4080

Now point your browser to http://localhost:4080 and login

Helm

Update Helm values located in values.yaml

Create the namespace:

$ kubectl create ns zinc

Install the chart:

$ helm install zinc helm/zinc -n zinc

Zinc can be available with an ingress or port-forward:

$ kubectl -n zinc port-forward svc/zinc 4080:4080

Data ingestion

Single record

curl example

$ curl
-u admin:Complexpass#123
-XPUT
-d '{"author":"Prabhat Sharma"}'
http://localhost:4080/api/myshinynewindex/document

Python example

import base64, json
import requests

user = "admin"
password = "Complexpass#123"
bas64encoded_creds = base64.b64encode(bytes(user + ":" + password, "utf-8")).decode("utf-8")


data = {
    "Athlete": "DEMTSCHENKO, Albert",
    "City": "Turin",
    "Country": "RUS",
    "Discipline": "Luge",
    "Event": "Singles",
    "Gender": "Men",
    "Medal": "Silver",
    "Season": "winter",
    "Sport": "Luge",
    "Year": 2006
  }

headers = {"Content-type": "application/json", "Authorization": "Basic " + bas64encoded_creds}
index = "games3"
zinc_host = "http://localhost:4080"
zinc_url = zinc_host + "/api/" + index + "/document"

res = requests.put(zinc_url, headers=headers, data=json.dumps(data))

Bulk ingestion

Bulk ingestion API follows same interface as Elasticsearch API defined in documentation.

curl -L https://github.com/prabhatsharma/zinc/releases/download/v0.1.1/olympics.ndjson.gz -o olympics.ndjson.gz

gzip -d olympics.ndjson.gz

curl http://localhost:4080/api/_bulk -i -u admin:Complexpass#123 --data-binary "@olympics.ndjson"

Search

Python example

import base64
import json
import requests

user = "admin"
password = "Complexpass#123"
bas64encoded_creds = base64.b64encode(
    bytes(user + ":" + password, "utf-8")).decode("utf-8")


params = {
    "search_type": "match",
    "query":
    {
        "term": "DEMTSCHENKO",
        "start_time": "2021-06-02T14:28:31.894Z",
        "end_time": "2021-12-02T15:28:31.894Z"
    },
    "from": 40, # use together with max_results for paginated results.
    "max_results": 20,
    "fields": ["_all"]
}

# params = {
#     "search_type": "querystring",
#     "query":
#     {
#         "term": "+City:Turin +Silver",
#         "start_time": "2021-06-02T14:28:31.894Z",
#         "end_time": "2021-12-02T15:28:31.894Z"
#     },
#     "fields": ["_all"]
# }

headers = {"Content-type": "application/json",
           "Authorization": "Basic " + bas64encoded_creds}
index = "games3"
zinc_host = "http://localhost:4080"
zinc_url = zinc_host + "/api/" + index + "/_search"

res = requests.post(zinc_url, headers=headers, data=json.dumps(params))

print(res.text)

Output

{
  "took": 0,
  "timed_out": false,
  "max_score": 7.6978611753656345,
  "hits": {
    "total": {
      "value": 3
    },
    "hits": [
      {
        "_index": "games3",
        "_type": "games3",
        "_id": "bd3e67f0-679b-4aa4-b0f5-81b9dc86a26a",
        "_score": 7.6978611753656345,
        "@timestamp": "2021-10-20T04:56:39.000871Z",
        "_source": {
          "Athlete": "DEMTSCHENKO, Albert",
          "City": "Turin",
          "Country": "RUS",
          "Discipline": "Luge",
          "Event": "Singles",
          "Gender": "Men",
          "Medal": "Silver",
          "Season": "winter",
          "Sport": "Luge",
          "Year": 2006
        }
      },
      {
        "_index": "games3",
        "_type": "games3",
        "_id": "230349d9-72b3-4225-bac7-a8ab31af046d",
        "_score": 7.6978611753656345,
        "@timestamp": "2021-10-20T04:56:39.215124Z",
        "_source": {
          "Athlete": "DEMTSCHENKO, Albert",
          "City": "Sochi",
          "Country": "RUS",
          "Discipline": "Luge",
          "Event": "Singles",
          "Gender": "Men",
          "Medal": "Silver",
          "Season": "winter",
          "Sport": "Luge",
          "Year": 2014
        }
      },
      {
        "_index": "games3",
        "_type": "games3",
        "_id": "338fea31-81f2-4b56-a096-b8294fb6cc92",
        "_score": 7.671309826309841,
        "@timestamp": "2021-10-20T04:56:39.215067Z",
        "_source": {
          "Athlete": "DEMTSCHENKO, Albert",
          "City": "Sochi",
          "Country": "RUS",
          "Discipline": "Luge",
          "Event": "Mixed Relay",
          "Gender": "Men",
          "Medal": "Silver",
          "Season": "winter",
          "Sport": "Luge",
          "Year": 2014
        }
      }
    ]
  },
  "buckets": null,
  "error": ""
}

API reference

These APIs can be used to programatically interact with Zinc.

All APIs must have an authorization header

e.g. Header:

Authorization: Basic YWRtaW46Q29tcGxleHBhc3MjMTIz

CreateIndex - Create a new index

Endpoint - PUT /api/index

While you do not need to create indexes manually as they are created automatically when you start ingesting the data, you could create them in advance using this API. S3 backed indexes must be created before they can be used.

e.g. PUT http://localhost:4080/api/index

Payload:

{ "name": "myshinynewindex", "storage_type": "s3" }

OR

{ "name": "myshinynewindex", "storage_type": "disk" }

Default storage_type is disk

DeleteIndex - Delete an index

Endpoint - DELETE /api/index/:indexName

This will delete the index and its associated metadata. Be careful using this as data is deleted unrecoverably.

e.g. DELETE http://localhost:4080/api/index/indextodelete

ListIndexes - List existing indexes

Endpoint - GET /api/index

Get the list of existing indexes

e.g. GET http://localhost:4080/api/index

UpdateMappings - Create/Update an index mappings

Endpoint - PUT /api/:target/_mappings

Create/Update a index mappings

e.g.

PUT http://localhost:4080/api/:target/_mappings

Payload:

{
	"mappings": {
		"properties": {
			"name": {
				"type": "text"
			},
			"author": {
				"type": "keyword"
			},
			"price": {
				"type": "numeric"
			},
			"published": {
				"type": "bool"
			},
			"create_at": {
				"type": "time"
			}
		}
	}
}

mappings also can defines when create index.

UpdateDocument - Create/Update a document and index it for searches

Endpoint - PUT /api/:target/document

Create/Update a document and index it for searches

e.g. PUT http://localhost:4080/api/myindex/document

Payload: { "name": "Prabhat Sharma" }

UpdateDocumentWithId - Create/Update a document and index it for searches. Provide a doc Id

Endpoint - PUT /api/:target/_doc/:id

Create/Update a document and index it for searches

e.g. PUT http://localhost:4080/api/myindex/_doc/1

Payload: { "name": "Prabhat Sharma is meeting friends in San Francisco" }

DeleteDocument - Delete a document

Endpoint - DELETE /api/:target/_doc/:id

This will delete the document from the index.

e.g. DELETE http://localhost:4080/api/myindex/_doc/1

Search - search for documents

Endpoint - POST /api/:target/_search

Search for documents

e.g. POST http://localhost:4080/api/stackoverflow-6/_search

Payload:

{
    "search_type": "matchphrase",
    "query": {
        "term": "shell window",
        "start_time": "2021-12-25T15:08:48.777Z",
        "end_time": "2021-12-28T16:08:48.777Z"
    },
    "sort_fields": ["-@timestamp"],
    "from": 1,
    "max_results": 20,
    "fields": [
        "*"
    ]
}

combine "from" and "max_results" to allow pagination.

sort_fields: list of fields to sort the results. Put a minus "-" before the field to change to descending order.

search_type can have following values:

  1. alldocuments
  2. wildcard
  3. fuzzy
  4. term
  5. daterange
  6. matchall
  7. match
  8. matchphrase
  9. multiphrase
  10. querystring

Search - aggregation

Endpoint - POST /api/:target/_search

Search for documents with aggregations

e.g. POST http://localhost:4080/api/stackoverflow-6/_search

Payload:

{
    "search_type": "matchall",
    "query": {
        "term": "query condition",
    },
    "aggs": {
      "my-aggs-term": {
        "agg_type": "terms",
        "field": "name",
        "size": 10
      }
    }
}

aggregation can filter data with query.

aggregation support:

  • Bucketing
    • Terms
    • Numeric Range
    • Date Range
  • Metrics
    • Min/Max/Count/Sum
    • Avg/Weighted Avg

BulkUpdate - Upload bulk data

Endpoint - POST /api/_bulk

This will upload multiple documents in batch. It is preferred over PUT /api/:target/_doc/:id when you have multiple documents to be inserted as it is magnitude of times faster than uploading individual documents.

You may want to experiment with actual number of documents that you send in a single request. An ideal number may range from 500 to 5000 documents. Log forwaders like fluentbit use this API.

e.g. POST /api/_bulk

Payload - ndjson (newline delimited json) content

e.g. ndjson contents

{ "index" : { "_index" : "olympics" } } 
{"Year": 1896, "City": "Athens", "Sport": "Aquatics", "Discipline": "Swimming", "Athlete": "HAJOS, Alfred", "Country": "HUN", "Gender": "Men", "Event": "100M Freestyle", "Medal": "Gold", "Season": "summer"}
{ "index" : { "_index" : "olympics" } } 
{"Year": 1896, "City": "Athens", "Sport": "Aquatics", "Discipline": "Swimming", "Athlete": "HERSCHMANN, Otto", "Country": "AUT", "Gender": "Men", "Event": "100M Freestyle", "Medal": "Silver", "Season": "summer"}
{ "index" : { "_index" : "olympics" } } 
{"Year": 1896, "City": "Athens", "Sport": "Aquatics", "Discipline": "Swimming", "Athlete": "DRIVAS, Dimitrios", "Country": "GRE", "Gender": "Men", "Event": "100M Freestyle For Sailors", "Medal": "Bronze", "Season": "summer"}
{ "index" : { "_index" : "olympics" } } 
{"Year": 1896, "City": "Athens", "Sport": "Aquatics", "Discipline": "Swimming", "Athlete": "MALOKINIS, Ioannis", "Country": "GRE", "Gender": "Men", "Event": "100M Freestyle For Sailors", "Medal": "Gold", "Season": "summer"}
{ "index" : { "_index" : "olympics" } } 
{"Year": 1896, "City": "Athens", "Sport": "Aquatics", "Discipline": "Swimming", "Athlete": "CHASAPIS, Spiridon", "Country": "GRE", "Gender": "Men", "Event": "100M Freestyle For Sailors", "Medal": "Silver", "Season": "summer"}

S3 storage (Experimental) for index data

Zinc can utilize s3 for storing index data. It still uses local disk for storing metadata. To enable storing data in an index you must do 2 things:

  1. Pass S3_BUCKET environment variable when zinc is starting up.
  2. Create an index before you use it with storage_type as s3.

e.g. PUT http://localhost:4080/api/index

Payload:

{ "name": "myobjectstoreindex", "storage_type": "s3" }

Zinc uses AWS SDK which looks for credentials through

  1. environment variables AWS_ACCESS_KEY_ID and AWS_SECRET_ACCESS_KEY
  2. Through your aws CLI credentials stored in ~/.aws/credentials
  3. Instance metadata - IMDS
  4. IAM Roles for service Accounts in EKS

Who uses Zinc (Known users)?

  1. Quadrantsec

Please do raise a PR adding your details if you are using Zinc.