With exempi 2.5.0 when the get_property was called on an unset one, it
crashed with "unkown error".
Fix with checking if property exists before actually getting the
property
Besides the exact GPS coordinate match and optional fuzzy match can be
used via the -d/--fuzzy-cache argument.
The default value is set to to 10m and can be set in the format number
and m/km. Eg 100m, 1km
During the run it calculates the distance from all previous cached
entries and if one falls inside the given fuzzy distance this lookup
data will be used and also stored in the direct GPS cache for further
direct GPS matches
Not only for path, but also for filename, country, state, city and
location.
There is a possibility that double byte characters are inside, so this
needs to be special formatted too
In some occasions Google might return non Latin based character strings
for reverse lookups.
Fixed with:
* add language=en to force english as primary return
* check each return set and only use entries that have Latin based
characters
When the path has double byte characters the length format does not work
and the path output will be too long for the space available.
We now use special double byte double count detection methods to adjust
the path length and output format
Fix almost all flake8 warnings for coding style
- long lines are currently ignored
- the Google lookup has been simplified
- the if and for loop are not yet simplified
If no exclude folders were given the list was not set so the initial
read failed.
Auto adjust to terminal size now works identical for smaller and larger
With flags -a to override any auto adjust and just keep default column
width. And flag -c for automatic compact view (40% column size
reduction)
The auto adjust on terminal size larger than needed is only for path
column at the moment.
For smaller it will auto adjust each column slowly in 10% steps until
the width is smaller than the terminal width.
Side note: CJK characters can still be reason for a skewed layout.
The include folders argument changed from -x to -i
-x is used for excluding folder or files form processing
In the list view the page number and max pages is also printed out
(cosmetic)
When we list data I strip out the path because if there are any CJK
files inside it will break the layout [TODO: fix this better]
The backup file creates a new backup file each time the file gets
written with ".BK.n." where n is a number starting from 1
Data was not written if some data was found in LR and then a Maps
lookup was done. This was because the overwrite check was done on set
data not original data
Add debug URL output to google (requests)
Add more valid types to look for an address in the returned data for
Google Maps API
Removed spaces from some lines
Skip XMP files with .BK. inside, they are backup files
Look into double-byte characters and print layout issues -> open
OpenStreetMap has optional parameter email to add an email address for
large requests amount.
Also check google and open street map options do not overlap.
All requests are now built with the parameters option when queries so
any parameter given is automatically correctly encoded.
If -s (--strict) argument is given it will check the path too and not
only the file basename.
The following bug fixes have been done:
- if verbose was not given it was unset and so google lookup debug
output crashed. Is set to 0 on unset
- if data is found multiple times in the Lightroom DB we print out a
different warning and not put it into the not found group
- after a google lookup write the data back into the base set for debug
output purposes
- if Lightroom data is used for overwrite data was never written because
the unset check was done on the already set data and not the original
unset data
