Dependece upon matrixprofile

[erezinman]

Hi,
It seems that the package matrixprofile is no longer maintained. The package forces a specific and deprecated version of protobuf, and hence is incompatible with many packages (tensorflow, pytorch, etc.). Please remove the dependence upon matrixprofile, or, at least, make this package's installation optional (maybe using extras_require in the setup.py?).

[arturdaraujo]

I fully agree. When creating an environment with pyarrow on python 3.7+ I had the same compatibility problem. Pyarrow is used for importing parquet and feather files which makes this incompatibility quite annoying for people who don't use csv.

(base) C:\Users\artur>conda create -n env_tsfresh36 python=3.9 spyder tsfresh=0.19 pyarrow
...
Collect all metadata (repodata.json): done
Solving environment: failed

LibMambaUnsatisfiableError: Encountered problems while solving:
  - package tsfresh-0.19.0-pyhd8ed1ab_0 requires matrixprofile >=1.1.10,<2.0.0, but none of the providers can be instald

[aiwalter]

Hi together, we are having same issue and would appreciate a solution here with a new tsfresh release!

[PJPRoche]

Hello, I am having the same issues with matrixprofile too. I raised an issue in their GitHub repo (matrix-profile-foundation/matrixprofile#102) with a suggestion to fix the issue, but there has not been any activity for a long time.
Would be great of the dependency can be removed here

[williamgilpin]

Hello, thank you very much for this incredible package. We are also having this issue, and have not yet been able to find a workaround.

How many of the features depend on matrixprofile? Would it be possible to drop those features for now (and thus the dependency) until an alternative like stumpy is implemented?

[arturdaraujo]

Any news on this issue @kempa-liehr ? This update would be incredibly helpful!

[enrique-wesper]

It would be amazing to be able to install tsfresh. I am using a mac M2 and I have also not been able to find a workaround

[kempa-liehr]

Fixed in https://github.com/blue-yonder/tsfresh/tree/v0.20.0.
Many thanks for your patience and perseverance.

Merry Christmas and a Happy New Year!