Jump to content

Alex

Members
  • Posts

    45
  • Joined

  • Last visited

  • Days Won

    1

Posts posted by Alex

  1. Hello,

    Please note that the NOT operator cannot be used without a positive assertion ('a NOT b', but not 'NOT b'). So the first example should not work.

    The second query is syntactically correct (though AND is redundant here). We need more information and possibly, sample data to be able to reproduce and diagnose the issue. I suggest to open a support ticket on the Intella support portal:  https://support.vound-software.com

     

  2. Hello,

    •  In the Content Analysis facet panel, create new category with a regular expression for your pattern ("New" button)
    • Execute content analysis for this category on the selected items ("Process->Content Analysis..." in the context menu)
    • Right-click the category name in the CA facet and select "Export values...". You can export the values list either into plain-text CSV or Excel format.
  3. Hello,

    To find these words in a sequence, a phrase pattern should be used:

    "(tr?s OR 3) (porcento OR percento) (jos? OR paulo)"

    Note that I omitted the clause with the singe % character, because it cannot be searched alone.

     

    Quote

    With this search, when I open an e-mail and click on Highlight hits

    Can you please elaborate a bit on what "click on Highlight hits" means? Intella UI does not have a button or other element named like this. The hits are highlighted automatically in the Previewer window when a keyword search result is opened.

  4. Hi,

    You can take a look at the regular expression-based Content Analysis feature that allows to search for complex patterns like in your case.

    The regular search does not support proximity searches with phrases and word distances. Only single terms and character distances are allowed:

    "Pace Slow"~20

    ("Pace" and "Slow" within 20 characters in the extracted text)

     

  5. On 1/18/2019 at 9:34 PM, cayerpm said:

    Searchable via RegEx or standard text search only?

    Either way. The search scope (text or metadata) is defined by selected search options only and does not depend on the syntax. Yes, regular expressions should be enclosed between forward slashes: otherwise, they are parsed just like normal query strings. 

    It should be noted that regex syntax support in the regular search is experimental and there are few known limitations related with it. For instance, only patterns for individual words are supported, so sentence or phrase search is not possible there.

    Regex search in the Content Analysis is a different story: it is more powerful, but as it was pointed out, works only on the full text. I think, Jon's reply was about this type of search.

     

  6. Hello,

    Metadata values should be searchable if the respective search options are turned on (the "Options" button under the search bar). 

    Regex search function in the Content Analysis ("custom categories") works only for full text of items (not metadata).

    If you think this is not working properly, feel free to open a support ticket on https://support.vound-software.com. Please attach the screenshots to illustrate the issue.

    Thanks!

  7. PDF's may contain photocopies or scanned images of documents. In this case, they need to be OCRed in order to be searchable in Intella. Please note that if embedded ABBYY FineReader method is used, all foreign languages must be selected on the configuration page of the OCR wizard.

    For PDF documents containing a parseable text, non-English language should not be a problem. Otherwise, we suggest to open a support ticket and attach few examples of problematic PDFs.  

     

  8. Dear users,

     

    Recently our support service received a number of complaints about Intella started to crash after updating the GeForce driver. As it turned out, the problem was the driver version, not compatible with Java applications.

     

    If your Intella installation is affected by this issue, it is highly recommended to download and install the 378.57 driver hot fix provided by NVIDIA:

     

    http://nvidia.custhelp.com/app/answers/detail/a_id/4378/

     

    • Like 1
  9. Hans,

     

    This is a good idea to have an option to export the search history to a KW list file. We will consider this in feature planning for future versions, 

     

    In the current version, it's possible to extract a list of all used keywords manually from "case.prefs" file located in "prefs" subfolder of the case. The entry to search in this file is "SearchHistory". The keywords in the list are separated by special "_." character sequences, which need to be removed and replaced by line breaks.

     

    Thanks!

  10. Hello Buldawg,

     

    You can use the Content Analysis facet that contains all Credit Cards and Social Security numbers found in this case. These values are extracted automatically during indexing, there is no need to run the Content Analysis process to identify them.

     

    The values can be exported to a CSV file using "Export values" option in the context menu.

     

    Hope this helps.

  11. Hello,

     

    Current implementation of the regular expression search allows to specify patterns for individual tokens (words), so it's like an advanced version of the wildcard syntax at the most. I doubt if it would help to execute complex proximity searches like in your example.

     

    We are currently working on an alternative implementation that will be a part of the Content Analysis functionality. This will allow to specify regular expressions that work on entire item content, without limitations of the token-based approach. Hopefully, this would solve the problem of complex proximity queries, among other things. This feature will be included into the next release.   

     

    Thanks for reporting the dead link, it will be removed from the next version of the User Manual.

     

    Good guides and tutorial on regular expression syntax can be found at http://regexone.com/.

  12. Hi Keaton

     

    One way to achieve this is to export an item ID list  by using "Export table as CSV" operation on the selected results with a single "Item ID" field. Then the ID list can be divided as needed using any text editor and the fragments are imported into "Item ID Lists" facet to create those item batches.

     

    Another way is to use the same procedure but with the "MD5 Hash" field and the "MD5 and Message Hash" facet. Unlike the first one, this method will ignore duplicate items (i.e. will work like the results table was deduplicated before exporting).

  13. Hi Keaton,

     

    > Is there a way to just reprocess the exception items?

     

    It's not possible in the current versions to re-process specific items in a selective way, so you would need either full reindexing or to add a new source containing corrected versions of the files as described by Adam.

     

    > When I have processing errors, will they be able to be resolved  by reprocessing?

     

    It depends on the nature of the error. For instance, if a corrupted source file is the problem, it is unlikely that repeated processing would help. From the other hand, some types of errors, like OutOfMemory can be resolved by tweaking memory parameters and reindexing.

     

    > How can I tell if a PDF was unable to be OCR'd? Will it be a Processing Error? I/O error?

     

    If an item is OCRed successfully and the OCRed text is imported back to Intella, the item should appear in the "OCRed" section of the Features facet. If this did not happen, it may indicate a problem with processing that item in the external OCR tool.

  14. Hi,

     

    > Is there a way to import a table containing various MD5 hash values and use it to match messages or attachments in an Intella case?

     

    You can import MD5 hash list into the "MD5 and Message Hash" facet panel for search. Please refer to the section 13.1.13 "MD5 and Message Hash" of Intella User Manual.

     

    > One is labeled "MD5 Hash" and the other is labeled "Message Hash." What is the difference, since both MD5 values ostensibly pertain to the same message?

     

    MD5 hash captures every detail of the message (including source-specific technical information) that is often not too useful for finding duplicate items. For instance, it is not uncommon if two copies of the same message get different MD5s if retrieved from different PST files. In contrast, Message Hashes are calculated only from message body and substantial headers, thus allowing loose message copies to be found. See the same section of the User Manual for detailed explanation between those types of hashes.

×
×
  • Create New...