Electronic document data and electronic device for viewing electronic document

Abstract

To provide an electronic device and an electronic document data displaying method which improve the visibility when a document having undergone electronification is viewed, and which improve the precision and speed of reading by a reader.SOLUTION: A method by an electronic device 10 for viewing an electronic document includes: identifying the segment of each sentence for a plurality of sentences forming a sequential line (1 line), and dividing the sentence at such a segment; justifying the header of each sentence at a left end; and performing a format change like high-lighting on a word so as to improve the visibility. The format change facilitates a user to overview the whole document, while easing a disadvantage such that the segment between the sentence becomes difficult to view. Furthermore, an effect of improving the visibility for the document written based on a regular rule improves.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present invention relates to an electronic device used to view electronic documents and a method for displaying electronic document data. [Background technology]

[0002] Compared to paper documents, electronic documents have many advantages, including reduced environmental impact, economical efficiency, and high accessibility. Furthermore, electronic documents offer the flexibility to change their display format to meet user needs. For example, when displaying program source code, it is common to use syntax highlighting to color-code the code based on grammar to improve its visibility. Web browsers also facilitate information retrieval by automatically highlighting and visually emphasizing words and phrases searched by users. Electronic documents also offer interactive features, such as automatically highlighting related text elements with clicks or taps, significantly improving document navigation. Furthermore, electronic documents can be integrated with computers, allowing users to, for example, select text elements they want to highlight, and even automatically calculate the most important text elements in a document. These customization features are described in detail in, for example, Japanese Patent Application Laid-Open Publication Nos. 2021-176068, 2021-043924, and 2021-206548, demonstrating the significant improvements in the convenience and functionality of electronic documents. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Japanese Patent Publication No. 2021-176068 [Patent Document 2] Japanese Patent Publication No. 2021-043924 [Patent Document 3] Japanese Patent Application Publication No. 2020-057337 [Patent Document 4] Patent Application No. 2021-206548 Summary of the Invention [Problem to be solved by the invention]

[0004] The present invention relates to prior art techniques for improving the visibility of electronic document usage, such as syntax highlighting and search term highlighting. These techniques focus on document components or specific words and visually highlight them to improve the visibility of the text. However, these methods are insufficient in improving visibility by taking into account the structure and context of the entire document. [Means for solving the problem]

[0005] The electronic device 10 for viewing electronic documents comprises: a character string data input device 12 for inputting first character string data 12x composed of a character group having display position information; a delimiter position determination device (14) that analyzes the first character string data (12x) and identifies delimiter positions (14x) of the sentence; a dividing device 16 that divides the first character string data 12x at the delimiter positions 14x to generate a second character string data group 16x; a line spacing setting device 18 for setting a first line spacing 18a; a display device (22) that displays the second character string data group (16x), the first character string data 12x and the second character string data group 16x may wrap around in the display area of ​​the display device 22; At the time of the folding, the first character string data 12x and each character string data in the second character string data group 16x are folded back with a second line spacing 18b relative to themselves, The first line spacing 18a is wider than the second line spacing 18b, the first character string data 12x includes a plurality of sentences, The second character string data group 16x is displayed on the display device 22 in a manner that the top edge is aligned if the first character string data 12x is displayed vertically, the left edge is aligned if the first character string data 12x is displayed horizontally, and the right edge is aligned if the first character string data 12x is displayed right. Electronic devices for viewing electronic documents10. [Effects of the Invention]

[0006] While conventional methods, such as highlighting and other formatting changes, make it easier to view a document from a bird's-eye view or reread it, have the problem of making sentence breaks difficult to see. The present invention solves this problem by aligning all sentence breaks at the left edge for horizontally written text (left-to-right binding), the right edge for horizontally written text (right-to-right binding), and the top edge for vertically written text. Furthermore, for documents written according to regular rules, the visibility improvement method provided by the present invention maximizes the effectiveness by making full use of those rules. This makes it easier to grasp the overall structure and context of a document at a glance, allowing readers to understand information more efficiently. [Brief explanation of the drawings]

[0007] [Figure 1] FIG. 1 is a diagram showing the components of the present invention together with the process flow of the present invention. [Figure 2] FIG. 2 shows another component of the present invention along with the process flow of the present invention. [Figure 3] FIG. 3 shows another component of the present invention along with the process flow of the present invention. [Figure 4] Figure 4 explains that the link part is a noun phrase. [Figure 5] FIG. 5 shows a document before the present invention is applied. [Figure 6] FIG. 6 shows the present invention applied to FIG. [Figure 7] Figure 7 shows Figure 6 with the paragraph background color changed. [Figure 8] Figure 8 shows Figure 5 with each word highlighted in a different color. [Figure 9]FIG. 9 shows the present invention applied to FIG. [Figure 10] Figure 10 shows the format of the noun phrases in Figure 5. [Figure 11] FIG. 11 shows the present invention applied to FIG. DETAILED DESCRIPTION OF THE INVENTION

[0008] In the present invention, a character string data input device 12 is first prepared to accept first character string data 12x provided by a user. This input device has the function of accepting a character group 12c including formatting information provided in formats such as HTML, PDF, e-book formats (e.g., EPUB, MOBI, AZW3), TeX, Markdown, and DOCX. This formatting information 12s includes elements that affect the reader's visibility, such as font type, bold, italic, underline, highlight, color, and character size, and is important for promoting the readability and comprehension of the document. For non-digitized documents or paper documents, optical character recognition (OCR) technology can be used to convert the data into electronic format and input it into the system. The character string data input device 12 is generally operated via a storage device (e.g., HDD or SSD) or a server, and data can be received using any device that provides character string data. The input character string data 12x serves to process and store the data up to the stage before the processing according to the present invention.

[0009] In the present invention, the first string data 12x may contain multiple sentences 12xs. It is important to first analyze this data using a sentence division position determination device 14 to identify the division positions 14x of each sentence. Specifically, in Japanese sentences, it is common to identify the division positions 14x of each sentence using periods such as ".", ".", ":," and ";." On the other hand, in English, segmentation is more complex, and it is necessary to consider cases where "." is used as a decimal point in numbers (e.g., 3.14). A simple and effective approach is to adopt a rule-based approach that considers a period followed by a space and then a capital letter to be the division position 14x of each sentence. It is also beneficial to introduce a machine learning task that assigns part-of-speech tags (POS tags) to periods for more accurate identification. For example, it is possible to distinguish between a dot indicating an abbreviation of a word and a period. Furthermore, it is also effective to use HTML layout information (e.g., p tag, h1 tag, h2 tag, h3 tag, etc.) and indentation associated with line breaks (start of paragraphs) to identify sentence break positions 14x. Through these analyses, the sentence segmentation device 16 appropriately segments the first character string data 12x and generates second character string data groups 16x.

[0010] In the present invention, it is also important that the first character string data 12x has display layout data 12z. In particular, in the case of HTML, data is managed by tags with structural information such as p and div, and styles such as CSS associated with them. As a specific example for explanation, consider a case where the first character string data 12x belongs to a p element (paragraph). In this case, the generated second character string data group 16x also belongs to the same p element, but line breaks are inserted between each data, and appropriate line spacing 18x is set by the line spacing setting device 18.

[0011] Three different settings are considered for the line spacing 18x. The first is the first line spacing 18a that is set when wrapping the text contained in the original p element. The second is the second line spacing 18b that is set between the data in the second string data group 16x, which should be set to the distance between the first line spacing 18a and the third line spacing 18c. The third is the third line spacing 18c that is set between two adjacent p elements (assuming there are two p elements). Considering semantic uniformity, the second line spacing 18b should be set at a longer distance than the first line spacing 18a, which is set for wrapping a single text contained in the same p element for display purposes, and should be set at a shorter distance than the third line spacing 18c, which is the line spacing between adjacent p elements.

[0012] The three line spacings 18x can be difficult to distinguish from one another. This often makes it difficult to see the difference between the second line spacing 18b and the third line spacing 18c. To solve this problem, it is effective to differentiate the first background color 20a of the p element from the second background color 20b of the third line spacing. The background color 20b is often the background color of the original document. Changing the background color 20 makes the difference between paragraphs more visually clear. Since the first background color 20a and the second background color 20b are usually the same, it is desirable to change them to a mild hue that does not impair visibility. Specifically, pastel colors are suitable. For example, when using black text, choosing a yellowish color for the first background color 20a or the second background color 20b is effective in improving readability.

[0013] As a final step, the modified second string data group 16x and display layout data 12z are input to the display device 22. This operation causes the reformatted electronic document to be displayed to the reader. The display device 22 may be a PC, tablet, smartphone, or the like. These devices typically include a graphics board and a display, enabling high-quality visual presentation of the document. In this way, the electronic document with improved visibility achieved by the present invention is finally presented to the user.

[0014] Furthermore, the present invention works in conjunction with a technology for changing the format information 12s of a part of a sentence, thereby enhancing the effect by making difficult documents easier to read. Below, we will show two examples of the technology for changing the format information 12s of a part of a sentence, and then explain how it works in conjunction with the present invention.

[0015] First, there is the technique of highlighting specific words. This technique is frequently used, particularly in the context of patent searches, to display search queries or predefined keywords in a specific color. This is done to identify where the search query or keywords are found. Another method is to calculate the importance of each word in a document and selectively highlight words based on that importance. This is advantageous in contexts where the user is unable to think of a search query or keywords, such as when the document is in an unfamiliar domain. It can also be combined with the previously mentioned method of user-specified keywords. This word highlighting technique highlights words that the reader considers important or that are interpreted as important by the document, thereby highlighting their overall importance while other words are locally important. This can be further emphasized by combining it with another technique described below. Word highlighting can also absorb differences in declension, for example, in inflected languages. The level of identity depends on the reader's perception of each word. Similarly, in space-separated languages, there are times when words should be considered subwords, such as when targeting one or both parts of a compound prefix and suffix. It is common in Nordic languages ​​such as German.

[0016] This is not necessary when determining whether to highlight based on an exact match, but tokenization by the tokenizer 24 is necessary when ambiguity is required as described above. New dynamic tokenization may be performed when a word to be highlighted is specified. This is because static tokenization may not highlight the word specified by the user. In the case of an exact match, dynamic tokenization is performed in a sense. This is because the sentence is divided, albeit locally, at the unit of the word to be highlighted. In summary, the tokenizer 24 obtains a third string data group 24x. The function data input device 26 inputs function data 26x that determines whether to apply the first format change 26sc to each of the third string data groups 24, and the calculation device 28 applies the function data 26x to the third string data groups 24x, obtaining a fourth string data group 28x that has been subjected to the first format change 26sc.

[0017] Another technique involves inputting the first string 12x into a parser 30, performing phrase structure analysis to compute a noun phrase 30np, and then applying a uniform second formatting change 30sc to the noun phrase 30np (e.g., displaying it in bold). This takes advantage of the human tendency to perceive things in a noun-based manner. For example, on a web page, formatting changes are applied to links, which are almost always noun phrases. See Figure 1, an excerpt from a Wikipedia article on graph neural networks (https: / / en.m.wikipedia.org / wiki / Graph_neural_network). Noun phrases are represented by (1) base nouns or (2) adjective-noun combinations, and this continues recursively. That is, adjective-(adjective-noun) is also a noun, and adjective-(adjective-(adjective-noun)) is also a noun. This chain continues, and the noun phrase it applies to is narrowed down by the addition of an adjective. In this way, there are multiple options for noun phrases, and one is selected based on multiple factors, such as whether it is semantically natural and whether it is long enough for the user to understand it. Languages ​​that use logographic characters (kanji) and do not have spaces as delimiters (Japanese, Chinese, etc.) have a larger amount of information per unit area than phonogram-based languages ​​(Indo-European languages ​​such as English), so it is advantageous to be able to use noun phrases with multiple options that contain a large amount of information.

[0018] These two formatting changes offer benefits such as a bird's-eye view of the document and ease of rereading. They are particularly effective in documents that emphasize accuracy or are explanatory, such as patent documents, legal documents, government publications, and research papers. From the perspective of accuracy, it is common for the same words to be repeated without omitting words, even if it makes sentences somewhat redundant. In such cases, the first technology can register or automatically calculate and highlight these repeated words, allowing for a clearer overview of the document while distinguishing between important and less important words. Furthermore, such documents are generally written with the expectation that they will be read back. For example, mathematical theorems, proofs, and contracts frequently use forward and backward references to terms. On the other hand, explanatory documents often distinguish between similar concepts, often using multiple adjectives to distinguish between similar concepts. In this case, using the second technique to separately format noun phrases (recursively composed of adjective-noun combinations) using boldface or similar not only makes them easier to read cognitively, but also allows the highlighted words in the noun phrase to be understood as the base noun (often the document's topic), with the unhighlighted adjective pair being understood as distinguishing the noun phrases that share the base. While the above explanation uses nouns as the base, adjectives can also serve as the base. The example discussed later also uses the latter. What's important is that one base is used and the other distinguishes concepts based on it, such as "A0 and B," "A1 and B," or "A and B0," "A and B1." In this case, combining the first technique (clarifying the base) and the second technique (clarifying the entire noun phrase) is effective in understanding documents that describe things.

[0019] Combining this invention with the two technologies mentioned above provides the following benefits. Each sentence can essentially be interpreted as a noun phrase. For example, a sentence like "I do this." can posit a hidden base noun, such as the fact that "I do this." Each sentence is a noun because it is recognized as a separate entity. Applying formatting techniques like bolding to each sentence makes the entire sentence bold, rendering it meaningless. Therefore, they should be separated using something other than bolding, typically a period (period). However, excessive use of the two technologies mentioned above increases the number of formatting changes, making the period less visible. Applying this invention to align the beginning of a sentence can compensate for the period's lack of visibility. It is best to complete each of the following three processes before processing the next: (1) determining the delimiter position, segmentation, and alignment for displaying alignment at the end of a sentence; (2) tokenizing for highlighting (if applicable); and (3) syntactic analysis and secondary formatting for noun phrase emphasis. However, each process can be processed independently, so the order in which they are applied is flexible. Since the layout change of the entire sentence due to alignment may cause anxiety for the reader, it is also effective to process each character string data before it is displayed on the display device. These methods make the structure of the entire document clear, allowing the reader to easily recognize the boundaries of each sentence. This is particularly effective in aiding the understanding of long documents or documents with complex structures. Next, a specific example in which the present invention is applied to a Japanese document will be shown. [Example]

[0020] FIG. 2 shows a portion of a pseudo-document created to facilitate explanation of the present invention. This document has the following characteristics. First, the document contains two paragraphs: "This document will discuss future directions as follows" and "Figure 1 shows graph construction and organization." Both are one-liners, that is, each consisting of a single string (first string data 10x). Second, these sentences are folded in the middle, and each line has a first line spacing 18a. Third, the two paragraphs have a third line spacing 18c.

[0021] First, we will explain the first embodiment (basic part) of the present invention. In Figure 3, a period "." is used to identify a delimiter position 14x, and a line break is performed at that position (second character string data group 16x). The second character string data groups are separated from each other by a second line spacing 18b, where the first line spacing 18a is closer than the second line spacing 18b, and the second line spacing 18b is closer than the third line spacing 18c. This line break allows each sentence to start on a new line and align with the left edge of the text. Each sentence in the first paragraph begins with "In Chapter XX," which provides a clearer overall view than a single sentence. Reorganizing the sentences in this way improves readability and unifies the overall appearance of the document. However, the introduction of a new line spacing 18, the second line spacing 18b, makes it difficult to distinguish between the three line spacings. In Figure 4, the backgrounds of the two paragraphs are changed, which improves the visibility of the third line spacing 18c.

[0022] Next, the present invention will be described in conjunction with formatting techniques.

[0023] First, we will explain the first form of highlighting words. Because patent documents cannot be displayed in color, the following figures enclose relevant sections in various brackets, such as " " and <>. However, we will assume that these sections are in different colors. (For those who would like to see the document in color, please refer to the inventor's website, functoy.com.) Bolding the highlighted sections is also effective, and we have used bold bands here. Figure 5 shows key words highlighted within a sentence. Specifically, "graph," "neural network," and "natural language processing" are highlighted in the same color as the "Chapter 1," "Chapter 2," and other phrases. While this allows users to easily see where each word is located and view the entire document from a bird's-eye view or reread it, it also makes it difficult to see where the sentences end. Figure 6 shows the results of combining highlighting with the first form of the present invention. As with the first form, the beginning of each sentence in the first paragraph begins with "In Chapter XX," providing a better overall view than Figure XX, which highlights a single sentence. On the other hand, when compared with Figure ○, which uses only the first format, it is possible to see that "Chapter ○" is at the beginning, and to get a bird's-eye view of the position of the main themes of "graphs," "neural networks," and "natural language processing" within the entire document. In this way, by combining this invention with formatting technology, important elements of the text are emphasized, making it possible to clarify the overall structure and promoting reader understanding.

[0024] Next, we will explain the case where noun phrase emphasis is added to the first form. Because it is obvious that the combination of highlighting and noun phrase emphasis reduces the visibility of punctuation marks, we will omit this explanation here. Below, we will assume the first form of the present invention and compare it with three other examples: (1) highlighting only (Figure 6, already mentioned), (2) noun phrase emphasis only (Figure 7), and (3) both highlighting and noun phrase emphasis (Figure 8). Figure 7 (noun phrase emphasis only) makes it easier to recognize chunks such as "graph-perspective natural language processing," "traditional graph-based methods," "graph construction," "graph representation learning," and "graph encoder-decoder model," thereby reducing the cognitive burden on the reader.

[0025] Figure 8 is Figure 7 with highlighting added. Please pay attention to the highlighted and unhighlighted parts within the noun phrases. Some obvious examples are "traditional graph-based methods" (traditional methods), "graph construction" (construction), "graph representation learning" (representation learning), and "graph encoder-decoder model" (encoder-decoder model). While the highlighted parts stand out, the unhighlighted parts () stand out conversely. This makes it immediately clear what the differences are between the chapters that also explain "graphs." These contrasts are not apparent in Figure 6, which has only highlighting and no boldface. Similarly, these contrasts are not apparent in Figure 7, which has only boldface and no highlighting.

[0026] Differences can be seen in "Applications," "Challenges," and "Future Directions" from Chapter 7 to Chapter 9. There is a well-defined rule: "In Chapter N, A is B," which makes this rule stand out when alignment is added with line breaks. Furthermore, A is a noun phrase, but here it is split due to its length, but it is "(adjective, highlighted) in graph neural network natural language processing" and "X (noun, unhighlighted)," and the unhighlighted X side stands out. Figure ○, which only has highlighting and noun phrase emphasis, does not benefit from this well-defined rule. [Explanation of symbols]

[0027] 10 Electronic devices for viewing electronic documents 12 First character string data input device 12x First string data 12p paragraph 12z display layout data 12s format information 14. Delimiter position determination device 14x separator points 16 Splitting device 16x Second string data group 18 Line spacing setting device 18x line spacing 18a First Interlinear 18b Second Interlinear 18c Third Interlinear 20a Primary background color 20b Secondary background color 22 Display device 22x viewing area 24 Tokenizer 24x Third string data group 26 Function data input device 26x Function Data 26sc First format change 28 Arithmetic unit 28x Fourth string data group 30 Syntax Parser 30x Fifth string data group 30np noun phrase 30sc Second format change

Claims

1. An electronic device for viewing electronic documents, a character string data input device for inputting first character string data composed of a character group having display position information; a delimiter position determination device that analyzes the first character string data and identifies sentence delimiter positions; a division device that divides the first character string data at the delimiter positions to generate a second character string data group; a line spacing setting device for setting a first line spacing; a display device that displays the second character string data group, the first character string data and the second character string data group may wrap around in a display area of ​​the display device; At the time of folding, the first character string data and each character string data in the second character string data group are folded with a second line spacing relative to themselves, the first line spacing is wider than the second line spacing; the first character string data includes a plurality of sentences, The second character string data group is displayed on the display device in a manner that the top edge is aligned if the first character string data is displayed vertically, the left edge is aligned if the first character string data is displayed horizontally, and the right edge is aligned if the first character string data is displayed right. Electronic devices for viewing electronic documents.

2. the first character string data has a plurality of paragraphs; a third space exists between the paragraphs; the first line spacing is wider than the second line spacing, and the first line spacing is narrower than the third line spacing; 2. The electronic device for viewing electronic documents according to claim 1.

3. the display device displays the background colors of the adjacent paragraphs and the background color of the third interline portion in different colors.

3. The electronic device for viewing electronic documents according to claim 2.

4. a tokenizer that tokenizes the first character string data or the second character string data and outputs third character string data divided into a certain unit; a function data input device for inputting function data having information on a conditional expression for determining whether or not to change the third character string data using the third character string data as an argument, and information on a format after the change; a calculation unit that applies the function data to the third character string data, applies a first format change to each character group, and outputs fourth character string data; The first to fourth character string data are composed of a character group having format information.

4. An electronic device for viewing electronic documents according to claim 1.

5. a parser that performs a syntactic analysis of the first character string data and outputs fifth character string data that is divided into units of noun phrases and others; the first to third character string data and the fifth character string data are composed of a character group having format information, the noun phrase in the fifth character string data is subjected to a second formatting change; the second formatting change is uniform within the noun phrase.

4. An electronic device for viewing electronic documents according to claim 1.

6. An electronic device for viewing electronic documents that satisfies both the conditions of claims 5 and 6, The first format change and the second format change are different from each other. Electronic devices for viewing electronic documents.

Images