fix for v9 SON files

[MarinManuel]

In the CED documentation for the SON library, it says:

In version 9 and above files, if pos is of type TDOF, the disk offset is pos*DISKBLOCK. In version 8 and previous files, the disk offset is pos.

this PR fixes the offsets for v9 files, and fixes #818

[samuelgarcia]

Hi @MarinManuel. This is super cool and should help many users.
Thanks a lot.

[samuelgarcia]

I have push the file "Two-mice-bigfile-test000.smr"
See https://gin.g-node.org/NeuralEnsemble/ephy_testing_data/src/master/spike2

Could add it here in neo/test/rawiotest and here in neo/test/iotest and check that test are passing ?

[JuliaSprenger]

The tests are failing as channel_names are expected to be of type 'S' and not unicode. @samuelgarcia was there a reason for this not to be unicode?
There are two options now:

1. change the definition of the ChannelIndex class and make channel_name unicode. This affects all ios based on the raw mechanism.
2. Hackily reading the channel_names as unicode and converting it to type S.

I would go for approach 2, since the ChannelIndex object will anyway be replaced in the future (#817), so we don't need to introduce unnecessary changes for ChannelIndex objects now.

[MarinManuel]

@JuliaSprenger @samuelgarcia I've added a line to convert back the unicode strings to ascii for the time being. Hopefully that does not introduce any more problems.

[samuelgarcia]

Hi all.
@JuliaSprenger this is strange. I thought we fix this old (and very anoying) behavior where everything was ascii string because of python 2. Now python27 is no more here and evrything should be unicode everwhere.

@MarinManuel : I propose to fix this str vs unicode problem outside this PR very very soon. So you could rebase this PR with the offcial fix. is it OK ?

[JuliaSprenger]

@MarinManuel the format change is in master now, could you rebase and undo your changes in basefromrawio.py?

[samuelgarcia]

Please remove this. this is now fixed.

[samuelgarcia]

Please remove this. this is now fixed.

[samuelgarcia]

@JuliaSprenger :Could we first merge this PR before #775 which is also related to spike2io ?

[JuliaSprenger]

Sure, then we keep #775 in the queue.

[samuelgarcia]

In fact I ask exactly the contary I wanted.
I wanted to say:
"Could we merge first #775 (mine older) and then #824 (this one)"

Sorry. One day I will read what I am writting before push the green button.

[MarinManuel]

@samuelgarcia Hi, I have reverted the commit on files channelindex.py and basefromrawio.py, is that good enough? Bit of a git newbie here...

[JuliaSprenger]

Hi @MarinManuel,
there are still conflicts between the current master branch on Neuralensemble and the master branch on your fork. What you can do is rebase your branch on top of the current Neuralensemble master. First make sure you also have an up-to-date local version of that one on your machine by running
git fetch upstream (assuming upstream points to Neuralensemble)
Then you can do the rebasing via
git rebase upstream/master - this will lead to the conflicts pointed out above to appear during the rebase in your local files. The conflicts are indicated in the files as

<<<<
first set of changes
=======
second set of changes
>>>>>

Please edit the files to include all the changes that are suggested by Neuralensemble master and additionally keep the ones you want to include in this PR.
Finally you add the files using git add <filename> and run git rebase --continue to finalize the rebase process.
To update this PR you then push the changes you made to your fork: git -f push origin master (assuming origin points to your fork). Here you need the -f flag as you want to overwrite the history of the commits in your fork, which were changed during the rebase process.

I hope this helps. Tell us if there's issues during the process.

[MarinManuel]

Hi @JuliaSprenger
Thank you for the detailed explanations. Hope this works now

[samuelgarcia]

Hi @MarinManuel
could you modify theses two files like this:

neo/test/rawiotest/test_spike2rawio.py

class TestSpike2RawIO(BaseTestRawIO, unittest.TestCase, ):
    rawioclass = Spike2RawIO
    files_to_download = [
        'File_spike2_1.smr',
        'File_spike2_2.smr',
        ...
        'Two-mice-bigfile-test000.smr', # <<<<<< new line
    ]
    entities_to_test = files_to_download

neo/test/iotest/test_spike2io.py

class TestSpike2IO(BaseTestIO, unittest.TestCase, ):
    ioclass = Spike2IO
    files_to_test = [
        'File_spike2_1.smr',
        'File_spike2_2.smr',
        ...
        'Two-mice-bigfile-test000.smr', # <<<<<< new line

And afte rthis I will be able to merge this PR.

thanks again.

[MarinManuel]

could you modify theses two files like this:
@samuelgarcia yes, done

[samuelgarcia]

thanks.
merging now.