declone and deskrule updates
Posted: 2023 Jan 02, 04:41
a big-ish update for declone that supports excluded folders, and a smaller one for deskrule
* EXCLUDED FOLDERS
You can specify folders to exclude from a search in the FOLDERS TO SEARCH box. Say that you want to scan all C:\ drive but not a particular folder C:\PRIVATE, you add both folders in the box, then you right click on the unwanted folder and use EXCLUDE THIS FOLDER menu command. Excluded folders appear in red letters. If you exclude a folder by accident, use the same menu command to turn it on.
To exclude folders with drag-drop, hold down <CTRL> key as you drop them into Folders To Search box.
Excluded folders must be subfolders of the main folder scanned, otherwise they are ignored.
Note that certain folders like C:\WINDOWS and other system and hidden folders are excluded from scans automatically (if SKIP PROTECTED FOLDERS advanced option is ticked)
* SIMILAR PHOTOS AND IMAGES
The AI for similar photos relies heavily on picture date (the date it was taken). A new property called AVERAGE BRIGHTNESS can be used to find generic similar images by content (not necessarily JPG photos), which is useful when "date taken" is irrelevant, e.g. comparing PNG or GIF images. You can add this property manually in advanced property page, or clear FILES MUST HAVE SAME EXTENSION tickbox and it will be selected automatically for you. This way you can discover photos saved in different formats, or/and resized. In this context it is recommended to keep the similarity percent to 95% or higher, to avoid false positives.
* BLURRED PHOTOS
An experimental image property called BLUR INDEX is added in Legacy category. This shouldn't be used as a scan property, but only as a quality indicator for cleanup. For example, after you search for similar photos, right click on the column header and use SELECT COLUMNS menu, then add the new BLUR property. Photos that are lower quality have high blur index (100 and above). So if you sort by this column, the first item in each group is the one with the "better quality", so if you use KEEP THE FIRST IN GROUP option in mark wizard, you get to keep the best photo of each bunch, and delete the rest.
This should be used with care because the AI is confused with photos of blue skies, or sea, as it interprets the uniformity as "bad quality" which isn't always the case.
* SET DEFAULT COLUMNS
Some file details like size and date modified appear in all scan results, whether you search for photos or e-books. If you want to see more or less columns, first set them up with SELECT COLUMNS (right click on the list header to see the menu), then use SET AS DEFAULT from the header menu. Note each scan category adds its own relevant properties automatically, e.g. dimensions for photos. So you should select NEUTRAL properties with this command that are relevant to all scan types, or at least for those you use more frequently.
* SIMILAR DOCUMENTS
Comparing similar text files saved as PDF/DOC by content gets easier and better with the new PLAIN NAME property (which is the file name without the extension). The same text document saved as DOC and PDF may have very different file size, but if the basic names are the same, they can be used to match duplicates, e.g. REPORT1.PDF and REPORT1.DOC. This new property is automatically selected if you untick SAME EXTENSION advanced property.
* STOCHASTIC REVISION
The AI algorithms that power some of the similarity scans are random in nature, so if you run them twice on the same files, you can get slightly different results. This is the case for similar photos and similar documents BY CONTENT. Therefore if you don't like some scan results, press CTRL+R keys and run the scan again without changing any of the parameters and see if you like the results better.
The REVISE command found under CLEAN UP toolbar button menu can be also used with modified properties if that makes sense for your scan. Revision uses the SAME files, and is quicker than a fresh new scan.
* FOCUSED DIALOGS. Whenever a dialog window or message appears, the main deskrule window is alpha-shaded, to help focusing on the message at hand. This behavior can be turned off setting the advanced registry option GAO2_NOSHADEDLG (=64)
* PRIMARY PROPERTY BOOST
Similar file detection is feasible only if a suitable master property is used to sort the files so that similar files are placed one next to another. Regardless of the similarity level you set, the first property (as it appears in ADVANCED property page) is matched at a higher level (95% by default); only the remaining properties are considered at the level you specify e.g. 90% or whatever.
A new registry tweak called n1stColBoost controls this extra percentage given to the primary column. However you are advised not to change it because if it is lowered, scan time will get much worse!
* In similarity scans you may want to split a large group that wasn't correctly identified, in 2 smaller groups. To achieve this, select the first item of the new subgroup then use SPLIT GROUP command from MARK DUPLICATES button drop-down menu.
* <Alt+Up> arrow keyboard shortcut jumps to the previous group (Down arrow was already available)