protoc-gen-druid-spec

Installation

go install github.com/nguyensinhtu/protoc-gen-druid-spec@latest

Usage

protoc --druid-spect_out=path/to/outdir foo.proto

protoc --druid-spec_out=path/to/out/dir foo.proto --proto_path=. --proto_path=<path_to_google_proto_folder>/src

Example

syntax = "proto3";

package foo;

import "druid_ingestion.proto";
import "druid_spec.proto";

import "google/protobuf/wrappers.proto";
import "google/protobuf/timestamp.proto";

message Bar {
  option (gen_druid_spec.druid_opts) = {
    data_source_name: "bar_proto3_table"
    segment_granularity: "day"
    query_granularity: "day"
  };

  string client_id = 1 [ (gen_druid_spec.spec) = {
    dimension : {
      multi_value_handling : "SORTED_SET"
      create_bitmap_index : {value: true}
    }
    metric : {
      metric_name : "client_id_sketch"
      size : 16384
      type : "thetaSketch"
    }
  } ];

  Baz baz = 2 
    [ (gen_druid_spec.spec).flatten = {prefix : "baz_" } ];;

  TimeSpec time_spec = 4[ (gen_druid_spec.spec).flatten = { prefix : "time_spec_" } ];
}

message Baz {
  int32 a = 1 [ (gen_druid_spec.spec).metric = {
    metric_name: "a_metric"
    size: 16384 
    type: "thetaSketch"
  }];
}

message TimeSpec {
  string date_key = 1[ (gen_druid_spec.spec).timestamp = {} ];
}

Output

{
 "spec": {
  "dataSchema": {
   "dataSource": "bar_proto3_table",
   "timestampSpec": {
    "column": "time_spec__date_key"
   },
   "dimensionsSpec": {
    "dimensions": [
     {
      "name": "client_id",
      "type": "string",
      "multiValueHandling": "SORTED_SET",
      "createBitmapIndex": true
     },
     {
      "name": "baz__a",
      "type": "long"
     }
    ],
    "dimensionExclusions": [
     "time_spec__date_key"
    ]
   },
   "metricsSpec": [
    {
     "name": "client_id_sketch",
     "type": "thetaSketch",
     "isInputThetaSketch": false,
     "fieldName": "client_id",
     "size": 16384
    },
    {
     "name": "a_metric",
     "type": "thetaSketch",
     "isInputThetaSketch": false,
     "fieldName": "baz__a",
     "size": 16384
    }
   ],
   "granularitySpec": {
    "type": "uniform",
    "segmentGranularity": "day",
    "queryGranularity": "day",
    "rollup": true,
    "intervals": []
   }
  },
  "ioConfig": {
   "inputFormat": {
    "type": "json",
    "flattenSpec": {
     "fields": [
      {
       "type": "jq",
       "name": "baz__a",
       "expr": ".baz.a"
      },
      {
       "type": "jq",
       "name": "time_spec__date_key",
       "expr": ".time_spec.date_key"
      }
     ],
     "useFieldDiscovery": false
    }
   }
  }
 }
}

Conceptually, after input data records are read, Druid applies ingestion spec components in a particular order: first flattenSpec (if any), then timestampSpec, then transformSpec, and finally dimensionsSpec and metricsSpec. Keep this in mind when writing your ingestion spec.

flattenSpec

Parent field separate with nested field by double underscore '__'

• if you dont set prefix or empty prefix, default will be protobuf field name

message Bar {
   Foo foo = 4[ (gen_druid_spec.spec).flatten = {} ];
}

message Foo {
  int32 i1 = 1;
}

Output

{
  "fields": [
    {
      "type": "jq",
      "name": "foo__i1",
      "expr": ".foo.i1"
    }
  ],
  "useFieldDiscovery": false
}

• if you want to remove parent name set ignore_name

message Bar {
   Foo foo = 4[ (gen_druid_spec.spec).flatten = {ignore_name: true} ];
}

message Foo {
  int32 i1 = 1;
}

Output

{
  "fields": [
    {
      "type": "jq",
      "name": "i1",
      "expr": ".foo.i1"
    }
  ],
  "useFieldDiscovery": false
}

• if you want to use diffirent prefix name

message Bar {
   Foo foo = 4[ (gen_druid_spec.spec).flatten = {prefix: "my_custom_name"} ];
}

message Foo {
  int32 i1 = 1;
}

Output

{
  "fields": [
    {
      "type": "jq",
      "name": "my_custom_name__i1",
      "expr": ".foo.i1"
    }
  ],
  "useFieldDiscovery": false
}

• if you want use diffirent name set output_name

message Bar {
   Foo foo = 4[ (gen_druid_spec.spec).flatten = {output_name: "rename_foo"} ];
}

message Foo {
  int32 i1 = 1;
}

Output

{
  "fields": [
    {
      "type": "jq",
      "name": "rename_foo__i1",
      "expr": ".rename.i1"
    }
  ],
  "useFieldDiscovery": false
}

timestampSpec

• if field is set to be timestamp it will be ignore in dimensionExclusions
• Can not apply flattenSpec, metricSpec, dimensionSpec with timestampSpec
• Can not apply timestampSpec to record type
• Can not apply multiple timestampSpec to one configured data source

Query filters

TODO - Support Selector filter, Logical expression filters

IO Config

We do not recommend config the IOConfig for specific schema, because we can re-use one schema for multiple input source , but if you want you still config it

message Bar {
  option (gen_druid_spec.druid_opts) = {
    data_source_name: "bar_proto3_table"
    
    io_config: {
      topic: "abc"
      bootstrap_servers: "PLAINTEXT://abc.com"
      use_earliest_offset: false
    }
    
    ingestion_type: "kafka"
  };
}

Output

{
 "spec": {
  "dataSchema": {
   "dataSource": "bar_proto3_table",
   "dimensionsSpec": {},
   "granularitySpec": {
    "type": "uniform",
    "segmentGranularity": "day",
    "queryGranularity": "none",
    "rollup": true,
    "intervals": []
   }
  },
  "ioConfig": {
   "consumerProperties": {
    "bootstrap.servers": "PLAINTEXT://abc.com"
   },
   "inputFormat": {
    "type": "json"
   },
   "topic": "abc",
   "useEarliestOffset": false
  }
 },
 "type": "kafka"
}