@platforma-open/milaboratories.software-ptabler.schema 1.11.2 → 1.11.3

package/dist/index.d.ts

@@ -1,10 +1,11 @@
-import { ReadCsvStep, ReadNdjsonStep, WriteCsvStep, WriteNdjsonStep, BaseFileReadStep, BaseFileWriteStep } from './io';
+import { ReadCsvStep, ReadNdjsonStep, WriteCsvStep, WriteNdjsonStep, BaseFileReadStep, BaseFileWriteStep, WriteParquetStep, ReadParquetStep } from './io';
 import { AddColumnsStep, FilterStep, SelectStep, WithColumnsStep, WithoutColumnsStep } from './basic_steps';
 import { AggregateStep } from './aggregate';
 import { AnyJoinStep } from './join';
 import { ConcatenateStep } from './concatenate';
 import { SortStep } from './sort';
-export type PTablerStep = ReadCsvStep | ReadNdjsonStep | WriteCsvStep | WriteNdjsonStep | AddColumnsStep | FilterStep | AggregateStep | AnyJoinStep | ConcatenateStep | SortStep | SelectStep | WithColumnsStep | WithoutColumnsStep;
+import { WriteFrameStep } from './write_frame';
+export type PTablerStep = ReadCsvStep | ReadNdjsonStep | ReadParquetStep | WriteCsvStep | WriteNdjsonStep | WriteParquetStep | AddColumnsStep | FilterStep | AggregateStep | AnyJoinStep | ConcatenateStep | SortStep | SelectStep | WithColumnsStep | WithoutColumnsStep | WriteFrameStep;
 export type PTablerWorkflow = {
     workflow: PTablerStep[];
 };

package/dist/io.d.ts

@@ -59,6 +59,11 @@ export interface ReadNdjsonStep extends BaseFileReadStep {
     /** The type of the step, which is always 'read_ndjson' for this operation. */
     type: 'read_ndjson';
 }
+/** Represents the configuration for a step that reads data from an Apache Parquet file into the tablespace. */
+export interface ReadParquetStep extends BaseFileReadStep {
+    /** The type of the step, which is always 'read_parquet' for this operation. */
+    type: 'read_parquet';
+}
 /**
  * Base interface for file writing operations that contains common fields
  * shared across different file format writers.
@@ -87,3 +92,10 @@ export interface WriteNdjsonStep extends BaseFileWriteStep {
     /** The type of the step, which is always 'write_ndjson' for this operation. */
     type: 'write_ndjson';
 }
+/**
+ * Represents the configuration for a step that writes a table from the tablespace to an Apache Parquet file.
+ */
+export interface WriteParquetStep extends BaseFileWriteStep {
+    /** The type of the step, which is always 'write_parquet' for this operation. */
+    type: 'write_parquet';
+}

package/dist/write_frame.d.ts

@@ -0,0 +1,57 @@
+/**
+ * Axes values are used as unique key to access the column data,
+ * so they cannot be floating point numbers.
+ */
+export type AxisType = 'Int' | 'Long' | 'String';
+/** PColumn values could be of any type supported by PFrames. */
+export type ColumnType = 'Int' | 'Long' | 'Float' | 'Double' | 'String';
+export interface AxisMapping {
+    /** The name of the column in the input table. */
+    column: string;
+    /**
+     * The type of the axis to be created.
+     * A cast would be performed if the type of the column in the input table is different.
+     * Warning: cast failure will result in an error.
+     */
+    type: AxisType;
+}
+export interface ColumnMapping {
+    /** The name of the column in the input table. */
+    column: string;
+    /**
+     * The value type of the PColumn to be created.
+     * A cast would be performed if the type of the column in the input table is different.
+     * Warning: cast failure will produce null values in the PColumn.
+     */
+    type: ColumnType;
+}
+/** Defines a step that creates a PFrame from the lazy table. */
+export interface WriteFrameStep {
+    /** The type identifier for this step. */
+    type: 'write_frame';
+    /** The name of the table from the tablespace which will be written. */
+    inputTable: string;
+    /**
+     * The name of the frame to be created.
+     * A folder with this name will be created in the working directory.
+     * The folder will contain the .datainfo and .parquet files representing the frame's PColumns.
+     */
+    frameName: string;
+    /**
+     * Axes specs, the order of axes in PColumns will match the order of axes in this array.
+     * Each column in the input table could be used as axis or PColumn only once.
+     * Warning: rows with null values in axes columns of input table will be discarded.
+     */
+    axes: AxisMapping[];
+    /**
+     * PColumns specs.
+     * Each column in the input table could be used as axis or PColumn only once.
+     */
+    columns: ColumnMapping[];
+    /**
+     * Partition key length, specifies the number of axes to be split into metadata partition key.
+     * Default: 0, which means no partitioning.
+     * Must be strictly less than the number of axes.
+     */
+    partitionKeyLength: number;
+}

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@platforma-open/milaboratories.software-ptabler.schema",
-  "version": "1.11.2",
+  "version": "1.11.3",
   "description": "Type definitions for PTabler",
   "types": "./dist/index.d.ts",
   "main": "./dist/index.js",
@@ -18,11 +18,11 @@
   ],
   "dependencies": {},
   "devDependencies": {
-    "@platforma-sdk/eslint-config": "",
     "vite-plugin-dts": "^4.5.3",
     "eslint": "^9.25.1",
     "typescript": "~5.6.3",
-    "vite": "^6.3.5"
+    "vite": "^6.3.5",
+    "@platforma-sdk/eslint-config": "1.1.0"
   },
   "scripts": {
     "type-check": "tsc --noEmit --composite false",

package/src/index.ts

@@ -1,15 +1,33 @@
-import type { ReadCsvStep, ReadNdjsonStep, WriteCsvStep, WriteNdjsonStep, BaseFileReadStep, BaseFileWriteStep } from './io';
-import type { AddColumnsStep, FilterStep, SelectStep, WithColumnsStep, WithoutColumnsStep } from './basic_steps';
+import type {
+  ReadCsvStep,
+  ReadNdjsonStep,
+  WriteCsvStep,
+  WriteNdjsonStep,
+  BaseFileReadStep,
+  BaseFileWriteStep,
+  WriteParquetStep,
+  ReadParquetStep,
+} from './io';
+import type {
+  AddColumnsStep,
+  FilterStep,
+  SelectStep,
+  WithColumnsStep,
+  WithoutColumnsStep,
+} from './basic_steps';
 import type { AggregateStep } from './aggregate';
 import type { AnyJoinStep } from './join';
 import type { ConcatenateStep } from './concatenate';
 import type { SortStep } from './sort';
+import type { WriteFrameStep } from './write_frame';
 
 export type PTablerStep =
   | ReadCsvStep
   | ReadNdjsonStep
+  | ReadParquetStep
   | WriteCsvStep
   | WriteNdjsonStep
+  | WriteParquetStep
   | AddColumnsStep
   | FilterStep
   | AggregateStep
@@ -18,7 +36,8 @@ export type PTablerStep =
   | SortStep
   | SelectStep
   | WithColumnsStep
-  | WithoutColumnsStep;
+  | WithoutColumnsStep
+  | WriteFrameStep;
 
 export type PTablerWorkflow = {
   workflow: PTablerStep[];

package/src/io.ts

@@ -64,6 +64,12 @@ export interface ReadNdjsonStep extends BaseFileReadStep {
   type: 'read_ndjson';
 }
 
+/** Represents the configuration for a step that reads data from an Apache Parquet file into the tablespace. */
+export interface ReadParquetStep extends BaseFileReadStep {
+  /** The type of the step, which is always 'read_parquet' for this operation. */
+  type: 'read_parquet';
+}
+
 /**
  * Base interface for file writing operations that contains common fields
  * shared across different file format writers.
@@ -104,3 +110,11 @@ export interface WriteNdjsonStep extends BaseFileWriteStep {
   /** The type of the step, which is always 'write_ndjson' for this operation. */
   type: 'write_ndjson';
 }
+
+/**
+ * Represents the configuration for a step that writes a table from the tablespace to an Apache Parquet file.
+ */
+export interface WriteParquetStep extends BaseFileWriteStep {
+  /** The type of the step, which is always 'write_parquet' for this operation. */
+  type: 'write_parquet';
+}

package/src/write_frame.ts

@@ -0,0 +1,66 @@
+/**
+ * Axes values are used as unique key to access the column data,
+ * so they cannot be floating point numbers.
+ */
+export type AxisType = 'Int' | 'Long' | 'String';
+
+/** PColumn values could be of any type supported by PFrames. */
+export type ColumnType = 'Int' | 'Long' | 'Float' | 'Double' | 'String';
+
+export interface AxisMapping {
+  /** The name of the column in the input table. */
+  column: string;
+  /**
+   * The type of the axis to be created.
+   * A cast would be performed if the type of the column in the input table is different.
+   * Warning: cast failure will result in an error.
+   */
+  type: AxisType;
+}
+
+export interface ColumnMapping {
+  /** The name of the column in the input table. */
+  column: string;
+  /**
+   * The value type of the PColumn to be created.
+   * A cast would be performed if the type of the column in the input table is different.
+   * Warning: cast failure will produce null values in the PColumn.
+   */
+  type: ColumnType;
+}
+
+/** Defines a step that creates a PFrame from the lazy table. */
+export interface WriteFrameStep {
+  /** The type identifier for this step. */
+  type: 'write_frame';
+
+  /** The name of the table from the tablespace which will be written. */
+  inputTable: string;
+
+  /**
+   * The name of the frame to be created.
+   * A folder with this name will be created in the working directory.
+   * The folder will contain the .datainfo and .parquet files representing the frame's PColumns.
+   */
+  frameName: string;
+
+  /**
+   * Axes specs, the order of axes in PColumns will match the order of axes in this array.
+   * Each column in the input table could be used as axis or PColumn only once.
+   * Warning: rows with null values in axes columns of input table will be discarded.
+   */
+  axes: AxisMapping[];
+
+  /**
+   * PColumns specs.
+   * Each column in the input table could be used as axis or PColumn only once.
+   */
+  columns: ColumnMapping[];
+
+  /**
+   * Partition key length, specifies the number of axes to be split into metadata partition key.
+   * Default: 0, which means no partitioning.
+   * Must be strictly less than the number of axes.
+   */
+  partitionKeyLength: number;
+}