[Plugin]E2m3u2bouquet

[Hannibal_Lecter]

UPDATE
1.9.0

1) Fixed bouquet customization mode using template. My mistake, please forgive me =) ... was “broken” while optimizing the m3u-parser code. It's all fixed, it's all checked out. Works as it should

thanks Dorik1972

[Hannibal_Lecter]

UPDATE
1.9.1

1) Fixed occurrence of 'nwame' error in DEBUG logging mode when using template mode

2) Unified the list of available folders for users to select/save files. The following folders are not available for users to choose: “/bin”, “/boot”, “/dev”, “/lib”, “/proc”, “/run”, “/sbin”, “/sys”, “/var”, “/usr”, “/picon”, “/piconlcd”. So you can't store your playlists in any of these folders, they won't be available for selection in the plugin menu

3) Small adjustments to the skins and display of various items in the plugin menu

4) Completely replaced and rewritten the code responsible for auto-selection of EPG “by similarity”. Now a faster mathematical algorithm for finding “similarity” is used (Jaro–Winkler similarity is used now instead of Damerau–Levenshtein distance), and you can specify the degree of search accuracy (not lower than the specified one) in percent in the playlist settings menu



Available range of similarity recognition accuracy from 75 to 100% (100% is an absolute match)

5) Minor changes in the code responsible for loading picons. The speed of processing and loading of picons is slightly increased.

6) Added translations of all locales due to the above changes

enigma2-plugin-extensions-e2m3u2bouquet_1.9.1_all
https://fastupload.io/df4e53863a7430e7

enigma2-plugin-extensions-e2m3u2bouquet_1.9.1_all.deb
https://fastupload.io/c010e364deb93b86

[Hannibal_Lecter]

UPDATE
1.9.2

1) Complete refactoring of the code responsible for auto-selecting descriptions from EPG.

2) Now you will have access to two of the fastest algorithms for finding “similarity” ... they are fundamentally different in terms of math, but give excellent results in terms of finding and evaluating “similarity”

a) - Jaro-Winkler - excellent and fast for short strings (channel names)

b) - TF-IDF - this “math” is what you encounter almost every day when you ask Uncle Google to find something you type into the search bar ... i.e. it is a basic algorithm that allows you to find documents (texts) that “that set of words” that you typed into the search bar is the “document/text” that you are looking for ... Works well and fast on large strings/texts (channel names)



c) auto - the plugin itself analyzes the length of the compared channel names and applies one of the available algorithms for each of the channel names

You can now combine “options” for auto-fit “similarity”, both in terms of speed and accuracy of results depending on your playlist and the EPG you are using

3) Small changes in EPG import algorithm
4) Updated all locales related to changes in the plugin menu


enigma2-plugin-extensions-e2m3u2bouquet_1.9.2_all.deb
enigma2-plugin-extensions-e2m3u2bouquet_1.9.2_all.ipk

Download
https://www.swisstransfer.com/d/12b6e39 ... 484ba5cdca

[Hannibal_Lecter]

UPDATE
1..9.3

1) Added support for Python 3.13-based images
2) Minor edits to the Greek locale

[Hannibal_Lecter]

UPDATE 1.9.4

1) A small fix to the regexp of the M3U parser.
2) In playlists settings added the ability to set the number of threads to process/download picons.



Previously, their number was statically set and applied to all playlists similarly. Now you can “play” with this parameter for each playlist to achieve the best results in the processing speed. Be careful when selecting the value of this parameter. Increasing the number of processing threads does not always lead to qualitative improvements in processing/downloading performance
3) System settings concerning IPTV have been added to the plugin settings. They will be displayed in the plugin settings menu only if they are supported by the Enigma2 image you are using



Even if in your Rnigma2 image, these settings are in the testing and experimentation phase and may not always work correctly. These settings have been added for the future
4) Small code optimization was performed

[Hannibal_Lecter]

UPDATE 1.9.5

1) Fixed crash when selecting the interval bouquet update method in plugin settings on any DreamOS-based image (my mistake, I accidentally “snagged” it when rearranging the code)

2) Fixed EPG preferred language selection menu in playlist settings for GP4.2 image for DM ONE and TWO (possibly also for any other E2-images in which country flag icons are set in SVG format)

3) In the preset filters selection menu for channel names the logic of impossibility of simultaneous selection of several filters is implemented

Simultaneously selecting two or more filters from a marked block defies any logic. Now it is impossible under any attempts of the user. Either one or none can be included.

4) Added check for python-multiprocessing module, if it is not present, it will be installed automatically. This module is used to implement multithreaded processing/loading of picons

[Hannibal_Lecter]

UPDATE
1.9.6

1) In the playlists settings added an on/off option to add a marker with the channel group name at the beginning of the userbouquet list. This function works only if “Multibouquet” mode is enabled



You can now add/exclude this marker as you see fit in userbouquets created by plugin

[Hannibal_Lecter]

UPDATE 1.9.7

1) Optimized algorithm of threaded picon loading/checking. Now it works even faster and with less RAM consumption

2) In the individual mode of channel sorting added the ability to select sorting type by “tvg-chno” tag values. Sometimes there are playlists where this tag is set and in its values, the channel number is specified (as it is considered necessary or correct to set it by IPTV service provider).

For example, the well-known PlutoTV



In the red block on the screenshot above, the sorting order is according to the numbers specified in the tag “tvg-chno”, in the green block - all other channels in the group are in the order as they are specified in the playlist (these channels do not have the tag “tvg-chno” or its value is not set)

3) Small changes to the log output code to make it more readable

ENJOY!

[Hannibal_Lecter]

UPDATE
1.9.8

1) Fixed crash when trying to open channel group selection/rearrangement menu with missing playlist file

2) Small changes in the logic of EPG auto-selection by channel name. Added logic for filtering “garbage” inserts at the beginning of channel names (e.g., [DE], [NL] or |FR|, DE| and other “abbreviations”).

3) Fixed builtins.KeyError error occurred when new broadcasts appeared in the provider's playlist when template files were used to form bouquets.

4) Reduced caching pause of channel_id->serviceref matching file data during EPG import.

5) Expanded list of supported file systems when auto-detecting mounted devices. This is used to determine devices (HDD/USB/SDCard) with the largest free space for storing temporary files when there is insufficient free RAM during plugin operation.

6) Revamped the logic for converting m3u playlist data to serviceref in userbouqets. Now multithreaded processing is used, which avoids the “wheel” busyness of the system at almost any playlist size and everything works faster.

thanks Dorik1972

[Hannibal_Lecter]

UPDATE
1.9.9

1) The playlist settings interface has been redesigned. Settings for EPG importing and downloading channel picons have been moved to separate menus





2) In the EPG import settings you now have the option to enable/disable certain fields. For example, if your XMLTV contains <credits> nodes, then by enabling the search for information about the creative team the corresponding information will be added to the program description. Information about the content rating of the program is taken from the <rating> nodes of the XMLTV (EPG) you are using, and if they are not there, the search will be performed in the name of the program and its description (some providers place it there as (16+), [18+], etc.). If you are using a skin that displays content rating information, it will be displayed. If you have a very old and “slow” receiver with weak RAM or internal flash - you can disable import descriptions of programs. This significantly speeds up the import process and saves hardware resources



3) Added multi-threaded processing of importing EPG events. Now you can set the number of processing threads in the settings. On DreamOS-based images, this significantly speeds up the process of importing data into the EPG SQL database. On all other images, you will get a 10-15% increase in EPG import speed compared to any of the previous versions of the plugin. Keep in mind that “more threads” doesn't always mean faster imports. Find the best result for each of the playlists you are using by experimentation

4) The creation of “serviceref” reworked is by the requirements of - ETSI EN 300 468

Code:
6.2.33 Service descriptor
Table 89: Service type coding

If you have picon downloading disabled, turn it on before generating bouquets. After generating bouquets you can disable it again

5) Cleaned plugin code, and removed some libraries and dependencies