@promptbook/pdf 0.100.1 โ†’ 0.100.3-0

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
@@ -1,7 +1,7 @@
1
1
  import type { string_agent_name } from '../../types/typeAliases';
2
2
  import type { string_url_image } from '../../types/typeAliases';
3
3
  import type { string_book } from './string_book';
4
- export interface AgentBasicInformation {
4
+ export type AgentBasicInformation = {
5
5
  /**
6
6
  * Name of the agent
7
7
  * This is the first line of the agent source
@@ -17,7 +17,7 @@ export interface AgentBasicInformation {
17
17
  * This is the line starting with "META IMAGE"
18
18
  */
19
19
  profileImageUrl: string_url_image;
20
- }
20
+ };
21
21
  /**
22
22
  * Parses basic information from agent source
23
23
  *
@@ -6,7 +6,7 @@ import type { AgentModelRequirements } from '../_misc/AgentModelRequirements';
6
6
  * - Creating regex patterns for parsing
7
7
  * - Applying its effects to agent model requirements
8
8
  */
9
- export interface CommitmentDefinition {
9
+ export type CommitmentDefinition = {
10
10
  /**
11
11
  * The type/name of this commitment (e.g., 'PERSONA', 'KNOWLEDGE', etc.)
12
12
  */
@@ -45,4 +45,4 @@ export interface CommitmentDefinition {
45
45
  * @returns Updated agent model requirements
46
46
  */
47
47
  applyToAgentModelRequirements(requirements: AgentModelRequirements, content: string): AgentModelRequirements;
48
- }
48
+ };
@@ -4,7 +4,7 @@ import type { TODO_any } from '../../../utils/organization/TODO_any';
4
4
  *
5
5
  * This is like "compiled" version of agent source
6
6
  */
7
- export interface AgentModelRequirements {
7
+ export type AgentModelRequirements = {
8
8
  /**
9
9
  * The system message that defines the agent's behavior and personality
10
10
  */
@@ -34,4 +34,4 @@ export interface AgentModelRequirements {
34
34
  * Each commitment can store its own data here
35
35
  */
36
36
  readonly metadata?: Record<string, TODO_any>;
37
- }
37
+ };
@@ -2,7 +2,7 @@ import type { ParsedCommitment } from './ParsedCommitment';
2
2
  /**
3
3
  * Result of parsing agent source for commitments
4
4
  */
5
- export interface AgentSourceParseResult {
5
+ export type AgentSourceParseResult = {
6
6
  /**
7
7
  * The agent name (first line)
8
8
  */
@@ -15,4 +15,4 @@ export interface AgentSourceParseResult {
15
15
  * Lines that are not commitments (for system message)
16
16
  */
17
17
  nonCommitmentLines: string[];
18
- }
18
+ };
@@ -2,7 +2,7 @@ import type { BookCommitment } from '../_base/BookCommitment';
2
2
  /**
3
3
  * Parsed commitment line from agent source
4
4
  */
5
- export interface ParsedCommitment {
5
+ export type ParsedCommitment = {
6
6
  /**
7
7
  * The commitment type (e.g., 'PERSONA', 'KNOWLEDGE')
8
8
  */
@@ -19,4 +19,4 @@ export interface ParsedCommitment {
19
19
  * Line number in the agent source (1-based)
20
20
  */
21
21
  lineNumber: number;
22
- }
22
+ };
@@ -5,7 +5,7 @@ import type { string_postprocessing_function_name } from '../../types/typeAliase
5
5
  /**
6
6
  * Options for validating a prompt result
7
7
  */
8
- export interface ValidatePromptResultOptions {
8
+ export type ValidatePromptResultOptions = {
9
9
  /**
10
10
  * The result string to validate
11
11
  */
@@ -23,11 +23,11 @@ export interface ValidatePromptResultOptions {
23
23
  * Note: This is for validation purposes only - postprocessing should be done before calling this function
24
24
  */
25
25
  postprocessingFunctionNames?: ReadonlyArray<string_postprocessing_function_name>;
26
- }
26
+ };
27
27
  /**
28
28
  * Result of prompt result validation
29
29
  */
30
- export interface ValidatePromptResultResult {
30
+ export type ValidatePromptResultResult = {
31
31
  /**
32
32
  * Whether the result is valid (passes all expectations and format checks)
33
33
  */
@@ -40,7 +40,7 @@ export interface ValidatePromptResultResult {
40
40
  * Error that occurred during validation, if any
41
41
  */
42
42
  error?: ExpectError;
43
- }
43
+ };
44
44
  /**
45
45
  * Validates a prompt result against expectations and format requirements.
46
46
  * This function provides a common abstraction for result validation that can be used
@@ -6,7 +6,7 @@ import type { Takeable } from './Takeable';
6
6
  * @deprecated [๐Ÿคก] Use some better functional library instead of `TakeChain`
7
7
  */
8
8
  export type WithTake<TValue extends Takeable> = TValue & ITakeChain<TValue>;
9
- export interface ITakeChain<TValue extends Takeable> {
9
+ export type ITakeChain<TValue extends Takeable> = {
10
10
  readonly value: TValue;
11
11
  then<TResultValue extends Takeable>(callback: (value: TValue) => TResultValue): WithTake<TResultValue>;
12
- }
12
+ };
@@ -1,7 +1,7 @@
1
1
  import type { string_filename } from '../../../types/typeAliases';
2
2
  import type { really_unknown } from '../../organization/really_unknown';
3
3
  /**
4
- * Tests if given string is valid URL.
4
+ * Tests if given string is valid file path.
5
5
  *
6
6
  * Note: This does not check if the file exists only if the path is valid
7
7
  * @public exported from `@promptbook/utils`
@@ -15,7 +15,7 @@ export declare const BOOK_LANGUAGE_VERSION: string_semantic_version;
15
15
  export declare const PROMPTBOOK_ENGINE_VERSION: string_promptbook_version;
16
16
  /**
17
17
  * Represents the version string of the Promptbook engine.
18
- * It follows semantic versioning (e.g., `0.100.0`).
18
+ * It follows semantic versioning (e.g., `0.100.2`).
19
19
  *
20
20
  * @generated
21
21
  */
@@ -9,12 +9,12 @@ import type { string_pipeline_url } from '../types/typeAliases';
9
9
  /**
10
10
  * Options for wizard methods
11
11
  */
12
- interface WizardOptions {
12
+ type WizardOptions = {
13
13
  /**
14
14
  * Whether to enable verbose logging
15
15
  */
16
16
  isVerbose?: boolean;
17
- }
17
+ };
18
18
  /**
19
19
  * Wizard for simple usage of the Promptbook
20
20
  * Look at `wizard` for more details
package/package.json CHANGED
@@ -1,6 +1,6 @@
1
1
  {
2
2
  "name": "@promptbook/pdf",
3
- "version": "0.100.1",
3
+ "version": "0.100.3-0",
4
4
  "description": "Promptbook: Run AI apps in plain human language across multiple models and platforms",
5
5
  "private": false,
6
6
  "sideEffects": false,
@@ -94,7 +94,7 @@
94
94
  "module": "./esm/index.es.js",
95
95
  "typings": "./esm/typings/src/_packages/pdf.index.d.ts",
96
96
  "peerDependencies": {
97
- "@promptbook/core": "0.100.1"
97
+ "@promptbook/core": "0.100.3-0"
98
98
  },
99
99
  "dependencies": {
100
100
  "crypto": "1.0.1",
package/umd/index.umd.js CHANGED
@@ -26,7 +26,7 @@
26
26
  * @generated
27
27
  * @see https://github.com/webgptorg/promptbook
28
28
  */
29
- const PROMPTBOOK_ENGINE_VERSION = '0.100.1';
29
+ const PROMPTBOOK_ENGINE_VERSION = '0.100.3-0';
30
30
  /**
31
31
  * TODO: string_promptbook_version should be constrained to the all versions of Promptbook engine
32
32
  * Note: [๐Ÿ’ž] Ignore a discrepancy between file name and entity name
@@ -721,7 +721,7 @@
721
721
  }
722
722
 
723
723
  /**
724
- * Tests if given string is valid URL.
724
+ * Tests if given string is valid file path.
725
725
  *
726
726
  * Note: This does not check if the file exists only if the path is valid
727
727
  * @public exported from `@promptbook/utils`
@@ -733,18 +733,25 @@
733
733
  if (filename.split('\n').length > 1) {
734
734
  return false;
735
735
  }
736
- if (filename.split(' ').length >
737
- 5 /* <- TODO: [๐Ÿง ][๐Ÿˆท] Make some better non-arbitrary way how to distinct filenames from informational texts */) {
736
+ // Normalize slashes early so heuristics can detect path-like inputs
737
+ const filenameSlashes = filename.replace(/\\/g, '/');
738
+ // Reject strings that look like sentences (informational text)
739
+ // Heuristic: contains multiple spaces and ends with a period, or contains typical sentence punctuation
740
+ // But skip this heuristic if the string looks like a path (contains '/' or starts with a drive letter)
741
+ if (filename.trim().length > 60 && // long enough to be a sentence
742
+ /[.!?]/.test(filename) && // contains sentence punctuation
743
+ filename.split(' ').length > 8 && // has many words
744
+ !/\/|^[A-Z]:/i.test(filenameSlashes) // do NOT treat as sentence if looks like a path
745
+ ) {
738
746
  return false;
739
747
  }
740
- const filenameSlashes = filename.split('\\').join('/');
741
748
  // Absolute Unix path: /hello.txt
742
749
  if (/^(\/)/i.test(filenameSlashes)) {
743
750
  // console.log(filename, 'Absolute Unix path: /hello.txt');
744
751
  return true;
745
752
  }
746
- // Absolute Windows path: /hello.txt
747
- if (/^([A-Z]{1,2}:\/?)\//i.test(filenameSlashes)) {
753
+ // Absolute Windows path: C:/ or C:\ (allow spaces and multiple dots in filename)
754
+ if (/^[A-Z]:\/.+$/i.test(filenameSlashes)) {
748
755
  // console.log(filename, 'Absolute Windows path: /hello.txt');
749
756
  return true;
750
757
  }
@@ -3639,9 +3646,15 @@
3639
3646
  }
3640
3647
  if (isValidUrl(knowledgeSourceContent)) {
3641
3648
  const url = knowledgeSourceContent;
3649
+ if (isVerbose) {
3650
+ console.info(`๐Ÿ“„ [1] "${name}" is available at "${url}"`);
3651
+ }
3642
3652
  const response = await fetch(url); // <- TODO: [๐Ÿง ] Scraping and fetch proxy
3643
3653
  const mimeType = ((_a = response.headers.get('content-type')) === null || _a === void 0 ? void 0 : _a.split(';')[0]) || 'text/html';
3644
3654
  if (tools.fs === undefined || !url.endsWith('.pdf' /* <- TODO: [๐Ÿ’ต] */)) {
3655
+ if (isVerbose) {
3656
+ console.info(`๐Ÿ“„ [2] "${name}" tools.fs is not available or URL is not a PDF.`);
3657
+ }
3645
3658
  return {
3646
3659
  source: name,
3647
3660
  filename: null,
@@ -3677,13 +3690,17 @@
3677
3690
  await tools.fs.mkdir(path.dirname(path.join(rootDirname, filepath)), { recursive: true });
3678
3691
  }
3679
3692
  catch (error) {
3693
+ if (isVerbose) {
3694
+ console.info(`๐Ÿ“„ [3] "${name}" error creating cache directory`);
3695
+ }
3680
3696
  // Note: If we can't create cache directory, we'll handle it when trying to write the file
3681
3697
  // This handles read-only filesystems, permission issues, and missing parent directories
3682
- if (error instanceof Error && (error.message.includes('EROFS') ||
3683
- error.message.includes('read-only') ||
3684
- error.message.includes('EACCES') ||
3685
- error.message.includes('EPERM') ||
3686
- error.message.includes('ENOENT'))) ;
3698
+ if (error instanceof Error &&
3699
+ (error.message.includes('EROFS') ||
3700
+ error.message.includes('read-only') ||
3701
+ error.message.includes('EACCES') ||
3702
+ error.message.includes('EPERM') ||
3703
+ error.message.includes('ENOENT'))) ;
3687
3704
  else {
3688
3705
  // Re-throw other unexpected errors
3689
3706
  throw error;
@@ -3698,13 +3715,17 @@
3698
3715
  await tools.fs.writeFile(path.join(rootDirname, filepath), fileContent);
3699
3716
  }
3700
3717
  catch (error) {
3718
+ if (isVerbose) {
3719
+ console.info(`๐Ÿ“„ [4] "${name}" error writing cache file`);
3720
+ }
3701
3721
  // Note: If we can't write to cache, we'll process the file directly from memory
3702
3722
  // This handles read-only filesystems like Vercel
3703
- if (error instanceof Error && (error.message.includes('EROFS') ||
3704
- error.message.includes('read-only') ||
3705
- error.message.includes('EACCES') ||
3706
- error.message.includes('EPERM') ||
3707
- error.message.includes('ENOENT'))) {
3723
+ if (error instanceof Error &&
3724
+ (error.message.includes('EROFS') ||
3725
+ error.message.includes('read-only') ||
3726
+ error.message.includes('EACCES') ||
3727
+ error.message.includes('EPERM') ||
3728
+ error.message.includes('ENOENT'))) {
3708
3729
  // Return a handler that works directly with the downloaded content
3709
3730
  return {
3710
3731
  source: name,
@@ -3726,6 +3747,9 @@
3726
3747
  }
3727
3748
  // TODO: [๐Ÿ’ต] Check the file security
3728
3749
  // TODO: [๐Ÿงน][๐Ÿง ] Delete the file after the scraping is done
3750
+ if (isVerbose) {
3751
+ console.info(`๐Ÿ“„ [5] "${name}" cached at "${path.join(rootDirname, filepath)}"`);
3752
+ }
3729
3753
  return makeKnowledgeSourceHandler({ name, knowledgeSourceContent: filepath }, tools, {
3730
3754
  ...options,
3731
3755
  rootDirname,
@@ -3740,7 +3764,12 @@
3740
3764
  throw new EnvironmentMismatchError('Can not import file knowledge in non-file pipeline');
3741
3765
  // <- TODO: [๐Ÿง ] What is the best error type here`
3742
3766
  }
3743
- const filename = path.join(rootDirname, knowledgeSourceContent).split('\\').join('/');
3767
+ const filename = path.isAbsolute(knowledgeSourceContent)
3768
+ ? knowledgeSourceContent
3769
+ : path.join(rootDirname, knowledgeSourceContent).split('\\').join('/');
3770
+ if (isVerbose) {
3771
+ console.info(`๐Ÿ“„ [6] "${name}" is a valid file "${filename}"`);
3772
+ }
3744
3773
  const fileExtension = getFileExtension(filename);
3745
3774
  const mimeType = extensionToMimeType(fileExtension || '');
3746
3775
  if (!(await isFileExisting(filename, tools.fs))) {
@@ -3782,6 +3811,12 @@
3782
3811
  };
3783
3812
  }
3784
3813
  else {
3814
+ if (isVerbose) {
3815
+ console.info(`๐Ÿ“„ [7] "${name}" is just a explicit string text with a knowledge source`);
3816
+ console.info('---');
3817
+ console.info(knowledgeSourceContent);
3818
+ console.info('---');
3819
+ }
3785
3820
  return {
3786
3821
  source: name,
3787
3822
  filename: null,