Path 1

import pandas as pd

import numpy as np

import os

import sys

import deepchem

from rdkit import Chem

from Bio.Data.IUPACData import unambiguous_dna_letters

from Bio.Alphabet import Alphabet

from Bio import SeqIO

def log(string, verbose=True):

  return seq_one_hot_encode(np.array(sequences))

# This could just be ambiguous_dna_letters, but that would be much higher dim.

class IUPACUnambiguousDNAWithN(Alphabet):

    letters = unambiguous_dna_letters + "N"

def encode_sequence_with_biopython(fname, file_type="fasta", alphabet=None):

    # np.ndarray: Shape (N_sequences, 4, sequence_length, 1).

    if alphabet is None:

        alphabet = IUPACUnambiguousDNAWithN()

    # TODO: if None, then get from the filename

    sequences = SeqIO.parse(fname, file_type, alphabet)

    # The label encoder is given characters for ACGTN

    letter_encoder = {l: i for i, l in enumerate(alphabet.letters)}

    alphabet_length = len(letter_encoder)

    # Peak at the first sequence to get the length of the sequence.

    first_seq = next(sequences)

    sequence_length = len(first_seq.seq)

    seqs = []

    seqs.append(_seq_to_encoded(first_seq, letter_encoder, alphabet_length, sequence_length))

    for other_seq in sequences:

      seqs.append(_seq_to_encoded(other_seq, letter_encoder, alphabet_length, sequence_length))

    # return np.expand_dims(np.array(seqs), -1)

    return np.array(seqs)

def _seq_to_encoded(seq, letter_encoder, alphabet_length, sequence_length):

    b = np.zeros((alphabet_length, sequence_length))

    seq_ints = [letter_encoder[s] for s in seq]

    b[seq_ints, np.arange(sequence_length)] = 1

    return b

def save_metadata(tasks, metadata_df, data_dir):

  """

>seq0

AC

>seq1

GA

    except ValueError:

      thrown = True

    assert thrown

  def test_encode_sequence_with_biopython(self):

      fname = "./data/example.fasta"

      encoded_seqs = dc.utils.save.encode_sequence_with_biopython(fname)

      expected = np.array([

        [[0, 0], [1, 0], [0, 0], [0, 1], [0, 0]],

        [[1, 0], [0, 1], [0, 0], [0, 0], [0, 0]],

      ])

      np.testing.assert_array_equal(expected, encoded_seqs)

conda install -y -q -c conda-forge pbr=3.1.1

conda install -y -q -c rdkit rdkit=2017.09.1

conda install -y -q -c conda-forge setuptools=39.0.1

conda install -y -q -c conda-forge biopython=1.71

yes | pip install $tensorflow==1.6.0