Testers required for OpenViX Python 3 images

[ccs]

What decides to use 562? It always used to be 506. Is it random?

I've just gone back to 5.4.012 (Release) on 2 different boxes, they both pick up 506 using frequency finder, as they always have done.

So 5.5.013.015.P3 does something different.

 

[ccs]

Is signal quality actually used? My Ultimo4K and ET10K both show 0xFFFF in the xml file for all mux's (except 562 on my ET10K which is a bit lower at 0xEB88).

Maybe a rounding error has crept into P3 ?

SNR's on both boxes show differences between mux's 506 and 562, much bigger on the ET10K, but none reach anything like 100%.

 

[Huevos]

Is signal quality actually used? My Ultimo4K and ET10K both show 0xFFFF in the xml file for all mux's (except 562 on my ET10K which is a bit lower at 0xEB88).

Maybe a rounding error has crept into P3 ?

SNR's on both boxes show differences between mux's 506 and 562, much bigger on the ET10K, but none reach anything like 100%.

First it reads the NIT, which takes time so the SNR reading has a chance to stabilize.
https://github.com/oe-alliance/Auto...uetsMaker/src/scanner/frequencyfinder.py#L394

SNR is a 16-bit int. Max value is 65535. If all the values are maxed out the end result is in the hands of the sort algorithm.
https://github.com/oe-alliance/Auto...uetsMaker/src/scanner/frequencyfinder.py#L553

Sort is done here:
https://github.com/oe-alliance/Auto...aker/src/scanner/frequencyfinder.py#L619-L622

 

[ccs]

First it reads the NIT, which takes time so the SNR reading has a chance to stabilize.
https://github.com/oe-alliance/Auto...uetsMaker/src/scanner/frequencyfinder.py#L394

SNR is a 16-bit int. Max value is 65535. If all the values are maxed out the end result is in the hands of the sort algorithm.
https://github.com/oe-alliance/Auto...uetsMaker/src/scanner/frequencyfinder.py#L553

Sort is done here:
https://github.com/oe-alliance/Auto...aker/src/scanner/frequencyfinder.py#L619-L622

In P3 it looks like a list is created in signal quality (low to high) and frequency (low to high) order.

The last system=0 entry in the list is then selected and as the qualities are always(?) the same, 562 gets the nod as it's the highest frequency.

In P2 (exactly the same code) 506 gets selected, neither the highest nor the lowest frequency, so the qualities must come into play, or P2/P3 gets it wrong.

 

[kiddac]

the vu zero 4k image is corrupt and not usable.

initrd_auto.bin

 

[twol]

the vu zero 4k image is corrupt and not usable.

initrd_auto.bin

building off my server at the moment - maybe available later tonight

 

[Huevos]

building off my server at the moment - maybe available later tonight

This is available for download now.

 

[BrokenUnusableAccount]

Probably this:

for number in customised["video"].keys():

Maybe needs to be this:

for number in list(customised["video"].keys()):

Yes, that's it.
Thank you.
And the result seems to be Python 2 compatible as well.

 

[ccs]

In P3 it looks like a list is created in signal quality (low to high) and frequency (low to high) order.

The last system=0 entry in the list is then selected and as the qualities are always(?) the same, 562 gets the nod as it's the highest frequency.

In P2 (exactly the same code) 506 gets selected, neither the highest nor the lowest frequency, so the qualities must come into play, or P2/P3 gets it wrong.

Not sure where this is going to be honest.

If my limited understanding is correct, P2 works (for me) because the mux selected happens to "work" but there doesn't appear to be any logic in the choice.

P3 doesn't work because the mux selected appears to follow the logic, but doesn't provide all the necessary data.

Maybe the lowest frequency with the best signal quality could be consider instead? That happens to work (for me), and is no less logical.

 

[Huevos]

Not sure where this is going to be honest.

If my limited understanding is correct, P2 works (for me) because the mux selected happens to "work" but there doesn't appear to be any logic in the choice.

P3 doesn't work because the mux selected appears to follow the logic, but doesn't provide all the necessary data.

Maybe the lowest frequency with the best signal quality could be consider instead? That happens to work (for me), and is no less logical.

2 choices:

• Edit the created xml file by hand to suit your needs (after all it does contain all the data).
• Find the mux with BBC ONE. If there are multiple instances choose the strongest. Means fiddling with a lot of code for little return.

 

[Huevos]

Im using Custom region but all channels are showing


View attachment 62223

It looks like it is ignoring the hacks to remove channels.

This should be fixed. I have rebuilt all the boxes on my list.

 

[dsayers]

This should be fixed. I have rebuilt all the boxes on my list.

I did a software update and noticed the bouquets are ok now.

I think tools.py needs updating for streams in custommix aswel. I was going to have a look later today.

Traceback (most recent call last):
< 30549.9415>   File "/usr/lib/enigma2/python/Plugins/SystemPlugins/AutoBouquetsMaker/scanner/main.py", line 494, in doBuildIndex
< 30549.9441>     self.manager.save(self.providers, self.dependents)
< 30549.9442>   File "/usr/lib/enigma2/python/Plugins/SystemPlugins/AutoBouquetsMaker/scanner/manager.py", line 155, in save
< 30549.9453>     services, providers[provider_key]["sections"] = Tools().customMix(self.services, provider_key, providers, self.providerConfigs[provider_key])
< 30549.9453>   File "/usr/lib/enigma2/python/Plugins/SystemPlugins/AutoBouquetsMaker/scanner/tools.py", line 194, in customMix
< 30549.9464>     if "%" not in url[:10]: # url not encoded
< 30549.9465> TypeError: a bytes-like object is required, not 'str'
< 30549.9466> [ePyObject] (CallObject(<bound method AutoBouquetsMaker.doBuildIndex of <class 'Plugins.SystemPlugins.AutoBouquetsMaker.scanner.main.AutoBouquetsMaker'>>,()) failed)

 

[Huevos]

I did a software update and noticed the bouquets are ok now.

I think tools.py needs updating for streams in custommix aswel. I was going to have a look later today.

Traceback (most recent call last):
< 30549.9415>   File "/usr/lib/enigma2/python/Plugins/SystemPlugins/AutoBouquetsMaker/scanner/main.py", line 494, in doBuildIndex
< 30549.9441>     self.manager.save(self.providers, self.dependents)
< 30549.9442>   File "/usr/lib/enigma2/python/Plugins/SystemPlugins/AutoBouquetsMaker/scanner/manager.py", line 155, in save
< 30549.9453>     services, providers[provider_key]["sections"] = Tools().customMix(self.services, provider_key, providers, self.providerConfigs[provider_key])
< 30549.9453>   File "/usr/lib/enigma2/python/Plugins/SystemPlugins/AutoBouquetsMaker/scanner/tools.py", line 194, in customMix
< 30549.9464>     if "%" not in url[:10]: # url not encoded
< 30549.9465> TypeError: a bytes-like object is required, not 'str'
< 30549.9466> [ePyObject] (CallObject(<bound method AutoBouquetsMaker.doBuildIndex of <class 'Plugins.SystemPlugins.AutoBouquetsMaker.scanner.main.AutoBouquetsMaker'>>,()) failed)

if "%" not in url[:10]: # url not encoded

Probably:

if b"%" not in url[:10]: # url not encoded

 

[dsayers]

if "%" not in url[:10]: # url not encoded

Probably:

if b"%" not in url[:10]: # url not encoded

Thats fixed it. Thanks

 

[white_westie]

A bit late to the party, but quick question - I have installed base image on a Gigablue UE UHD, should I do a software update first before doing any testing?

 

[ccs]

2 choices:

• Edit the created xml file by hand to suit your needs (after all it does contain all the data).
• Find the mux with BBC ONE. If there are multiple instances choose the strongest. Means fiddling with a lot of code for little return.

I thought I'd found a workaround, but ran into problems.

If I'm changing the contents of terrestrial_finder.xml, what's the exact process I need to follow, the obvious one sometimes produces unexplained results.

Does, for example, frequency finder delete files when creating terrestrial_finder.xml - eg providers.cache?

What I found was that going back to my "working" version didn't find more than 50% of channels, whatever I tried.

 

[BrokenUnusableAccount]

After a couple of initial hiccups, which were largely of my own making, this is working superbly well on my Zgemma H7S.
I think it's actually a tiny bit faster than the normal Py2 image too.

 

[twol]

A bit late to the party, but quick question - I have installed base image on a Gigablue UE UHD, should I do a software update first before doing any testing?

I have just updated feeds for the gbue4k so might be worth while…….but I would make sure you have set everything up first and backed up your settings

 

[white_westie]

I have just updated feeds for the gbue4k so might be worth while…….but I would make sure you have set everything up first and backed up your settings

Okay 157 updated packages, so I assume you have been busy.....

Configured everything from scratch as I normally do, and so far have only had 1 crash

Menu -> Information -> Devices
Displays 3 lines for each of the device types, then crashes


Meant to add, my non standard OSCAM working fine at the moment!

 

[Huevos]

I thought I'd found a workaround, but ran into problems.

If I'm changing the contents of terrestrial_finder.xml, what's the exact process I need to follow, the obvious one sometimes produces unexplained results.

Does, for example, frequency finder delete files when creating terrestrial_finder.xml - eg providers.cache?

What I found was that going back to my "working" version didn't find more than 50% of channels, whatever I tried.

It just makes an xml file. Nothing more.