[rank] adds path count table for ranking (sourcegraph#47994)

Adds table to count the global paths that will be used in the ranking
background jobs.

## Test plan
Tested manually.

<!-- All pull requests REQUIRE a test plan:
https://docs.sourcegraph.com/dev/background-information/testing_principles
-->

internal/database/schema.json

@@ -467,6 +467,15 @@
  "Increment": 1,
  "CycleOption": "NO"
  },
+ {
+ "Name": "codeintel_ranking_path_counts_inputs_id_seq",
+ "TypeName": "bigint",
+ "StartValue": 1,
+ "MinimumValue": 1,
+ "MaximumValue": 9223372036854775807,
+ "Increment": 1,
+ "CycleOption": "NO"
+ },
  {
  "Name": "codeintel_ranking_references_id_seq",
  "TypeName": "bigint",
@@ -7576,6 +7585,114 @@
  ],
  "Triggers": []
  },
+ {
+ "Name": "codeintel_ranking_path_counts_inputs",
+ "Comment": "",
+ "Columns": [
+ {
+ "Name": "count",
+ "Index": 4,
+ "TypeName": "integer",
+ "IsNullable": false,
+ "Default": "",
+ "CharacterMaximumLength": 0,
+ "IsIdentity": false,
+ "IdentityGeneration": "",
+ "IsGenerated": "NEVER",
+ "GenerationExpression": "",
+ "Comment": ""
+ },
+ {
+ "Name": "document_path",
+ "Index": 3,
+ "TypeName": "text",
+ "IsNullable": false,
+ "Default": "",
+ "CharacterMaximumLength": 0,
+ "IsIdentity": false,
+ "IdentityGeneration": "",
+ "IsGenerated": "NEVER",
+ "GenerationExpression": "",
+ "Comment": ""
+ },
+ {
+ "Name": "graph_key",
+ "Index": 5,
+ "TypeName": "text",
+ "IsNullable": false,
+ "Default": "",
+ "CharacterMaximumLength": 0,
+ "IsIdentity": false,
+ "IdentityGeneration": "",
+ "IsGenerated": "NEVER",
+ "GenerationExpression": "",
+ "Comment": ""
+ },
+ {
+ "Name": "id",
+ "Index": 1,
+ "TypeName": "bigint",
+ "IsNullable": false,
+ "Default": "nextval('codeintel_ranking_path_counts_inputs_id_seq'::regclass)",
+ "CharacterMaximumLength": 0,
+ "IsIdentity": false,
+ "IdentityGeneration": "",
+ "IsGenerated": "NEVER",
+ "GenerationExpression": "",
+ "Comment": ""
+ },
+ {
+ "Name": "processed",
+ "Index": 6,
+ "TypeName": "boolean",
+ "IsNullable": false,
+ "Default": "false",
+ "CharacterMaximumLength": 0,
+ "IsIdentity": false,
+ "IdentityGeneration": "",
+ "IsGenerated": "NEVER",
+ "GenerationExpression": "",
+ "Comment": ""
+ },
+ {
+ "Name": "repository",
+ "Index": 2,
+ "TypeName": "text",
+ "IsNullable": false,
+ "Default": "",
+ "CharacterMaximumLength": 0,
+ "IsIdentity": false,
+ "IdentityGeneration": "",
+ "IsGenerated": "NEVER",
+ "GenerationExpression": "",
+ "Comment": ""
+ }
+ ],
+ "Indexes": [
+ {
+ "Name": "codeintel_ranking_path_counts_inputs_pkey",
+ "IsPrimaryKey": true,
+ "IsUnique": true,
+ "IsExclusion": false,
+ "IsDeferrable": false,
+ "IndexDefinition": "CREATE UNIQUE INDEX codeintel_ranking_path_counts_inputs_pkey ON codeintel_ranking_path_counts_inputs USING btree (id)",
+ "ConstraintType": "p",
+ "ConstraintDefinition": "PRIMARY KEY (id)"
+ },
+ {
+ "Name": "codeintel_ranking_path_counts_inputs_graph_key_and_repository",
+ "IsPrimaryKey": false,
+ "IsUnique": false,
+ "IsExclusion": false,
+ "IsDeferrable": false,
+ "IndexDefinition": "CREATE INDEX codeintel_ranking_path_counts_inputs_graph_key_and_repository ON codeintel_ranking_path_counts_inputs USING btree (graph_key, repository)",
+ "ConstraintType": "",
+ "ConstraintDefinition": ""
+ }
+ ],
+ "Constraints": null,
+ "Triggers": []
+ },
  {
  "Name": "codeintel_ranking_references",
  "Comment": "References for a given upload proceduced by background job consuming SCIP indexes.",

internal/database/schema.md

@@ -972,6 +972,22 @@ Foreign-key constraints:
 
 ```
 
+# Table "public.codeintel_ranking_path_counts_inputs"
+```
+ Column | Type | Collation | Nullable | Default 
+---------------+---------+-----------+----------+------------------------------------------------------------------
+ id | bigint | | not null | nextval('codeintel_ranking_path_counts_inputs_id_seq'::regclass)
+ repository | text | | not null | 
+ document_path | text | | not null | 
+ count | integer | | not null | 
+ graph_key | text | | not null | 
+ processed | boolean | | not null | false
+Indexes:
+ "codeintel_ranking_path_counts_inputs_pkey" PRIMARY KEY, btree (id)
+ "codeintel_ranking_path_counts_inputs_graph_key_and_repository" btree (graph_key, repository)
+
+```
+
 # Table "public.codeintel_ranking_references"
 ```
  Column | Type | Collation | Nullable | Default 

migrations/frontend/1677008591_add_codeintel_rank_path_counts_inputs/down.sql

@@ -0,0 +1 @@
+DROP TABLE IF EXISTS codeintel_ranking_path_counts_inputs;

migrations/frontend/1677008591_add_codeintel_rank_path_counts_inputs/metadata.yaml

@@ -0,0 +1,2 @@
+name: add_codeintel_rank_path_counts_inputs
+parents: [1677005673]

migrations/frontend/1677008591_add_codeintel_rank_path_counts_inputs/up.sql

@@ -0,0 +1,10 @@
+CREATE TABLE IF NOT EXISTS codeintel_ranking_path_counts_inputs (
+ id bigserial PRIMARY KEY NOT NULL,
+ repository text NOT NULL,
+ document_path text NOT NULL,
+ count int NOT NULL,
+ graph_key text NOT NULL,
+ processed boolean NOT NULL DEFAULT false
+);
+
+CREATE INDEX IF NOT EXISTS codeintel_ranking_path_counts_inputs_graph_key_and_repository ON codeintel_ranking_path_counts_inputs (graph_key, repository);

migrations/frontend/squashed.sql

@@ -1729,6 +1729,24 @@ CREATE SEQUENCE codeintel_ranking_exports_id_seq
 
 ALTER SEQUENCE codeintel_ranking_exports_id_seq OWNED BY codeintel_ranking_exports.id;
 
+CREATE TABLE codeintel_ranking_path_counts_inputs (
+ id bigint NOT NULL,
+ repository text NOT NULL,
+ document_path text NOT NULL,
+ count integer NOT NULL,
+ graph_key text NOT NULL,
+ processed boolean DEFAULT false NOT NULL
+);
+
+CREATE SEQUENCE codeintel_ranking_path_counts_inputs_id_seq
+ START WITH 1
+ INCREMENT BY 1
+ NO MINVALUE
+ NO MAXVALUE
+ CACHE 1;
+
+ALTER SEQUENCE codeintel_ranking_path_counts_inputs_id_seq OWNED BY codeintel_ranking_path_counts_inputs.id;
+
 CREATE TABLE codeintel_ranking_references (
  id bigint NOT NULL,
  upload_id integer NOT NULL,
@@ -4187,6 +4205,8 @@ ALTER TABLE ONLY codeintel_ranking_definitions ALTER COLUMN id SET DEFAULT nextv
 
 ALTER TABLE ONLY codeintel_ranking_exports ALTER COLUMN id SET DEFAULT nextval('codeintel_ranking_exports_id_seq'::regclass);
 
+ALTER TABLE ONLY codeintel_ranking_path_counts_inputs ALTER COLUMN id SET DEFAULT nextval('codeintel_ranking_path_counts_inputs_id_seq'::regclass);
+
 ALTER TABLE ONLY codeintel_ranking_references ALTER COLUMN id SET DEFAULT nextval('codeintel_ranking_references_id_seq'::regclass);
 
 ALTER TABLE ONLY configuration_policies_audit_logs ALTER COLUMN sequence SET DEFAULT nextval('configuration_policies_audit_logs_seq'::regclass);
@@ -4425,6 +4445,9 @@ ALTER TABLE ONLY codeintel_ranking_definitions
 ALTER TABLE ONLY codeintel_ranking_exports
  ADD CONSTRAINT codeintel_ranking_exports_pkey PRIMARY KEY (id);
 
+ALTER TABLE ONLY codeintel_ranking_path_counts_inputs
+ ADD CONSTRAINT codeintel_ranking_path_counts_inputs_pkey PRIMARY KEY (id);
+
 ALTER TABLE ONLY codeintel_ranking_references
  ADD CONSTRAINT codeintel_ranking_references_pkey PRIMARY KEY (id);
 
@@ -4857,6 +4880,8 @@ CREATE INDEX codeintel_ranking_definitions_upload_id ON codeintel_ranking_defini
 
 CREATE UNIQUE INDEX codeintel_ranking_exports_graph_key_upload_id ON codeintel_ranking_exports USING btree (graph_key, upload_id);
 
+CREATE INDEX codeintel_ranking_path_counts_inputs_graph_key_and_repository ON codeintel_ranking_path_counts_inputs USING btree (graph_key, repository);
+
 CREATE INDEX codeintel_ranking_references_upload_id ON codeintel_ranking_references USING btree (upload_id);
 
 CREATE INDEX configuration_policies_audit_logs_policy_id ON configuration_policies_audit_logs USING btree (policy_id);