@pipedog/core 0.0.0 → 0.2.0

package/dist/utils/text/base_text_filter.js

@@ -0,0 +1,50 @@
+// Copyright (c) 2024-2026, Daily
+// Copyright (c) 2026, pipedog contributors
+// SPDX-License-Identifier: BSD-2-Clause
+//
+// Ported from pipecat: src/pipecat/utils/text/base_text_filter.py
+/**
+ * Base text filter interface for Pipecat text processing.
+ *
+ * This module defines the abstract base class for text filters that can
+ * modify text content in the processing pipeline, including support for
+ * settings updates and interruption handling.
+ */
+/**
+ * Abstract base class for text filters in the Pipecat framework.
+ *
+ * Text filters are responsible for modifying text content as it flows through
+ * the processing pipeline. They support dynamic settings updates and can
+ * handle interruptions to reset their internal state.
+ *
+ * Subclasses must implement all abstract methods to define specific filtering
+ * behavior, settings management, and interruption handling logic.
+ */
+export class BaseTextFilter {
+    // Accepted (and ignored) so subclasses can forward `**kwargs` as upstream does.
+    constructor(_options = {}) { }
+    /**
+     * Update the filter's configuration settings.
+     *
+     * Subclasses should implement this method to handle dynamic configuration
+     * updates during runtime, updating internal state as needed.
+     *
+     * @param settings Dictionary of setting names to values for configuration.
+     */
+    async updateSettings(_settings) { }
+    /**
+     * Handle interruption events in the processing pipeline.
+     *
+     * Subclasses should implement this method to reset internal state, clear
+     * buffers, or perform other cleanup when an interruption occurs.
+     */
+    async handleInterruption() { }
+    /**
+     * Reset the filter state after an interruption has been handled.
+     *
+     * Subclasses should implement this method to restore the filter to normal
+     * operation after an interruption has been processed and resolved.
+     */
+    async resetInterruption() { }
+}
+//# sourceMappingURL=base_text_filter.js.map

package/dist/utils/text/index.d.ts

@@ -0,0 +1,2 @@
+export {};
+//# sourceMappingURL=index.d.ts.map

package/dist/utils/text/index.js

@@ -0,0 +1,7 @@
+// Copyright (c) 2024-2026, Daily
+// Copyright (c) 2026, pipedog contributors
+// SPDX-License-Identifier: BSD-2-Clause
+//
+// Ported from pipecat: src/pipecat/utils/text/__init__.py
+export {};
+//# sourceMappingURL=index.js.map

package/dist/utils/text/markdown_text_filter.d.ts

@@ -0,0 +1,147 @@
+/**
+ * Markdown text filter for removing Markdown formatting from text.
+ *
+ * This module provides a text filter that converts Markdown content to plain
+ * text while preserving structure and handling special cases like code blocks
+ * and tables.
+ *
+ * DIVERGENCE: upstream converts Markdown to HTML with the Python-Markdown
+ * package. Core takes no runtime dependencies, so `markdownToHtml` below is a
+ * small built-in converter that follows Python-Markdown's rules for the
+ * constructs that matter here (paragraphs, headers, rules, lists, quotes,
+ * indented code, tables, code spans, links, images, emphasis, escapes and raw
+ * HTML). Exotic Markdown (reference links, nested lists) may convert
+ * differently.
+ */
+import { BaseTextFilter } from "./base_text_filter.ts";
+/** Configuration parameters for Markdown text filtering. */
+export interface MarkdownTextFilterInputParamsOptions {
+    /** Whether to apply Markdown filtering. Defaults to true. */
+    enableTextFilter?: boolean | null;
+    /** Whether to remove code blocks from the text. Defaults to false. */
+    filterCode?: boolean | null;
+    /** Whether to remove table content from the text. Defaults to false. */
+    filterTables?: boolean | null;
+    /** Whether to remove repeated sequences from the text. Defaults to true. */
+    filterRepeatedSequences?: boolean | null;
+}
+/**
+ * Configuration parameters for Markdown text filtering
+ * (`MarkdownTextFilter.InputParams` upstream).
+ *
+ * - `enableTextFilter`: whether to apply Markdown filtering. Defaults to true.
+ * - `filterCode`: whether to remove code blocks from the text. Defaults to false.
+ * - `filterTables`: whether to remove table content from the text. Defaults to false.
+ * - `filterRepeatedSequences`: whether to remove repeated sequences from the
+ *   text. Defaults to true.
+ */
+export declare class MarkdownTextFilterInputParams {
+    enableTextFilter: boolean | null;
+    filterCode: boolean | null;
+    filterTables: boolean | null;
+    filterRepeatedSequences: boolean | null;
+    constructor(options?: MarkdownTextFilterInputParamsOptions);
+}
+/** Options for `MarkdownTextFilter`. */
+export interface MarkdownTextFilterOptions {
+    /** Configuration parameters for filtering behavior. */
+    params?: MarkdownTextFilterInputParams | null;
+}
+/**
+ * Text filter that removes Markdown formatting from text content.
+ *
+ * Converts Markdown to plain text while preserving the overall structure,
+ * including leading and trailing spaces. Handles special cases like asterisks
+ * and table formatting. Supports selective filtering of code blocks and
+ * tables based on configuration.
+ */
+export declare class MarkdownTextFilter extends BaseTextFilter {
+    /** Configuration parameters for Markdown text filtering. */
+    static InputParams: typeof MarkdownTextFilterInputParams;
+    protected _settings: MarkdownTextFilterInputParams;
+    protected _inCodeBlock: boolean;
+    protected _inTable: boolean;
+    protected _interrupted: boolean;
+    /**
+     * Initialize the Markdown text filter.
+     *
+     * @param options.params Configuration parameters for filtering behavior.
+     */
+    constructor(options?: MarkdownTextFilterOptions);
+    /**
+     * Update the filter's configuration settings.
+     *
+     * Keys may be given as upstream writes them (`enable_text_filter`) or in
+     * camelCase (`enableTextFilter`).
+     *
+     * @param settings Dictionary of setting names to values for configuration.
+     */
+    updateSettings(settings: Readonly<Record<string, unknown>>): Promise<void>;
+    /**
+     * Apply Markdown filtering transformations to the input text.
+     *
+     * @param text The input text containing Markdown formatting to be filtered.
+     * @returns The filtered text with Markdown formatting removed or converted.
+     */
+    filter(text: string): Promise<string>;
+    /**
+     * Handle interruption events in the processing pipeline.
+     *
+     * Resets the filter state and clears any tracking variables for code
+     * blocks and tables.
+     */
+    handleInterruption(): Promise<void>;
+    /**
+     * Reset the filter state after an interruption has been handled.
+     *
+     * Clears the interrupted flag to restore normal operation.
+     */
+    resetInterruption(): Promise<void>;
+    /**
+     * Remove code blocks from the input text.
+     *
+     * Handles interruptions and delegates to specific methods based on the
+     * current state.
+     */
+    protected _removeCodeBlocks(text: string): string;
+    /**
+     * Handle text when currently inside a code block.
+     *
+     * If we find the end of the block, return text after it. Otherwise, skip
+     * the content.
+     */
+    protected _handleInCodeBlock(match: {
+        start: number;
+        end: number;
+    } | null, text: string): string;
+    /** Handle text when not currently inside a code block. */
+    protected _handleNotInCodeBlock(match: {
+        start: number;
+        end: number;
+    } | null, text: string, codeBlockPattern: string): string;
+    /**
+     * Handle the case where a code block starts.
+     *
+     * Return any text before the code block and set the state to inside a code
+     * block.
+     */
+    protected _handleStartOfCodeBlock(text: string, startIndex: number): string;
+    /**
+     * Handle code blocks found within text content.
+     *
+     * If it's a complete code block, remove it and return surrounding text. If
+     * it's the start of a code block, return text before it and set state.
+     */
+    protected _handleCodeBlockWithinText(text: string, codeBlockPattern: string): string;
+    /**
+     * Remove HTML tables from the input text.
+     *
+     * Handles cases where both start and end tags are in the same input, as
+     * well as tables that span multiple text chunks.
+     *
+     * @param text The text containing HTML tables to remove.
+     * @returns The text with tables removed.
+     */
+    removeTables(text: string): string;
+}
+//# sourceMappingURL=markdown_text_filter.d.ts.map