Fix reading blackrock digital events.

[histed]

Blackrock specs say packet_insertion_reason is
1 for PARALLEL and 129 for SERIAL. (also 64 for PERIODIC, not implemented)
i.e. bit 1 is set for parallel and bits 1 and 7 are set for serial.

Because the bitfield is just a unique ID, I removed the bit testing and use equality instead.

---

Apologies: I didn't write tests for this code. But this fixes a bug where both parallel and serial events were being returned as parallel due to the bit testing above.

[pep8speaks]

Hello @histed! Thanks for submitting the PR.

• In the file neo/rawio/blackrockrawio.py, following are the PEP8 issues :

Line 347:30: E127 continuation line over-indented for visual indent
Line 889:25: E127 continuation line over-indented for visual indent
Line 1699:28: E127 continuation line over-indented for visual indent
Line 1700:28: E127 continuation line over-indented for visual indent
Line 1857:66: E231 missing whitespace after ','
Line 1858:100: E501 line too long (134 > 99 characters)

[coveralls]

Coverage decreased (-6.0%) to 49.944% when pulling 591e156 on histedlab:mh/blackrock_digital into 7753fb2 on NeuralEnsemble:master.

[coveralls]

Coverage decreased (-6.0%) to 49.944% when pulling 591e156 on histedlab:mh/blackrock_digital into 7753fb2 on NeuralEnsemble:master.

[samuelgarcia]

Thank you for the patch.
I look good.
I don't have read blacrock spec.
I let @JuliaSprenger @bjoern1001001 or @mdenker check this.
Best

[muellerbjoern]

Hi @histed,

thank you for your contribution.

The Blackrock manual that I have available is quite ambiguous regarding this issue.
I talked to @JuliaSprenger and we both think what you did makes more sense than the current implementation.

I would only suggest a small change: Instead of raising an error when an unknown event code is found, a warning would fit better. Then the user knows there might be more data in the file that cannot be processed, but the rest of the content remains available.

[JuliaSprenger]

I agree the manual is ambiguous and @histed interpretation makes more sense. @histed: Could you change the ValueError to a warning?