2

In the movie The Matrix, the character Neo is using a program that downloads and searches (newspaper) articles.

The tool has a GUI and it displays and scrolls the document it is currently searching. When the end of this document is reached, it downloads/shows the next one. This process seems to go off automatically, as Neo is sleeping meanwhile.

Neo sleeps while a tool on his computer downloads and searches newspaper articles.

A closer look at the GUI shows a toolbar, a download progress bar (top right), and the document area (which scrolls down automatically) with the status label "Searching…".

GUI of Neo’s searching tool

Is there a (similar) tool like this?

A simple keyword search is sufficient.

The tool must show and automatically scroll the document it is currently searching. As this would probably take less than a second per document, it should slow down somewhat, so that human observers can follow the search.

Bonus points for highlighting sentences/paragraphs the tool is currently searching. Even more bonus points for using OCR.

It would be great if the tool really downloads documents from the Internet (e.g., using Wget in the background), but searching local documents only is okay, too.

I don’t want to give file format requirements, but it should at least support a document type that allows to embed images and/or supports formatted text. HTML and PDF would be great.

I don’t care about platform/price/license.

  • Fun question. How would you specify which documents to look into? List of folder/filepaths? Note that search engine don't go through the entire documents one by one when answering a query but use some indices. – Franck Dernoncourt Jun 24 '14 at 15:04
  • 1
    @FranckDernoncourt: If it searches local documents, specifying a folder and/or a file name pattern (for a global search) would be nice; if it searches online, specifying a list of URLs and/or using a "follow hyperlinks" feature (starting from a given URL) would be nice (i.e., like Wget does it). -- But I don’t want to require a specific way, as I fear there won’t be many (if any at all?) tools that use such an ineffectual way of searching. – unor Jun 24 '14 at 15:12
  • @unor I have to ask: for what are you going to use it? – RockPaperLizard May 21 '15 at 20:34
  • 1
    @RockPaperLizard: I hope it will be relaxing and fun to watch a search, similar to some mesmerizing screensavers (with the added benefit that it’s actually useful, and that there’s a chance of serendipitous discoveries as it shows you all specified documents, i.e., even those that are not relevant results). – unor May 21 '15 at 23:02
  • See also on Movies & TV: What program is Neo using? – unor Feb 2 at 8:06
4

This is:

  • A command line solution
  • Works only for text files
  • Works for Linux (and perhaps on Windows, via Cygwin)

    while read line; do sleep 0.2s; echo "$line" | sed -r 's/word_you_want_to_find/\x1b[31m\1\x1b[0m/g'; done < <(find . -type f -exec cat {} \; )
    

... will do what you want.

  • You can control the speed by modifying the interval of sleep (I've used 0.2 seconds).
  • You can search for multiple words by using (word1|word2|word3) in place of word_you_want_to_find.
  • You can specify filename patterns too, for example, by using find . -type f -name '*.txt' -exec cat {} \;.
2

I have developed a web application like this, it is basically a web crawler that shows the relevant pages that it finds during it's crawl. You search by entering keywords, giving them weights and set a minimum relevance score for what websites that should be shown.

All the websites are shown in iframes so you can fully interact with them meanwhile the search goes on.

The source code can be found here: https://github.com/spydon/fetchur

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.