Special characters in txt output files

[ikajic]

If there are special characters (e.g. < or >) in the output txt files in data folder, pd.DataFrame(pytry.read('data')) will fail to read the data.

One use-case where this happened is when using nengo_ocl and recording the context in the output file. This line causes the problem:
context = <pyopencl.Context at 0x2afa190 on <pyopencl.Device 'GeForce GTX 970' on 'NVIDIA CUDA' at 0x28f8d80>>