如何使用python有条件地从txt文件中删除行序列

Question

我从MS-DIAL metabolomics MSP spectral kit containing EI-MS, MS/MS下载了一个大文本文件

文件以 txt 文件形式打开，如下所示：

NAME: C11H11NO5; PlaSMA ID-967
PRECURSORMZ: 238.0712
PRECURSORTYPE: [M+H]+
FORMULA: C11H11NO5
Ontology: Formula predicted
INCHIKEY:
SMILES:
RETENTIONTIME: 1.74
CCS: -1
IONMODE: Positive
COLLISIONENERGY:
Comment: Annotation level-3; PlaSMA ID-967; ID title-AC_Bulb_Pos-629; Max plant tissue-LE_Ripe_Pos
Num Peaks: 2
192.06602   53
238.0757    31

NAME: Malvidin-3,5-di-O-glucoside; PlaSMA ID-3141
PRECURSORMZ: 656.19415
PRECURSORTYPE: [M+H]+
FORMULA: C29H35O17
Ontology: Anthocyanidin O-glycosides
INCHIKEY: CILLXFBAACIQNS-UHFFFAOYNA-O
SMILES: COC1=CC(=CC(OC)=C1O)C1=C(OC2OC(CO)C(O)C(O)C2O)C=C2C(OC3OC(CO)C(O)C(O)C3O)=CC(O)=CC2=[O+]1
RETENTIONTIME: 2.81
CCS: 241.3010517
IONMODE: Positive
COLLISIONENERGY:
Comment: Annotation level-1; PlaSMA ID-3141; ID title-Malvidin-3,5-di-O-glucoside; Max plant tissue-Standard only
Num Peaks: 0

每个化合物在NAME 到下一个NAME 之间都有数据。

我正在尝试做的是删除所有在Num Peaks: 中的值为零的化合物（即Num Peaks: 0。如果化合物的第 12 行是Num Peaks: 0，则删除 Thins 化合物的所有数据 - 12 行向上，删除）。

在上面的复合中，删除NAME: Malvidin-3,5-di-O-glucoside; PlaSMA ID-3141到Num Peaks: 0之间的行 之后，我需要将数据保存回txt或msp格式。

我所做的只是将数据作为列表导入：

with open('path\to\MSMS-Public-Pos-VS15.msp') as f:
    lines = f.readlines()

然后创建一个带有索引的列表，其中每个化合物以 link 开头：

indices = [i for i, s in enumerate(lines) if 'NAME' in s]

我想，现在我需要附加差异大于 14 的连续索引（意味着峰值数大于零）link

# to find the difference between consecutive indices.

v = np.diff(indices)

选择那些有差异的 14 并在第一个位置添加一个元素 0

diff14 = np.where(v == 14)

diff14 = np.append([0],diff14[0])

现在我只想选择那些不在 diff14 中的值，以便创建一个包含峰数大于零的化合物的新列表

现在我需要一些循环来选择正确的索引，但不知道如何：

lines[indices[diff14[0]]: indices[diff14[1]]]

lines[indices[diff14[1]+1] : indices[diff14[2]]]

lines[indices[diff14[2]+1] : lines[indices[diff14[3]]]]

lines[indices[diff14[3]+1] : indices[diff14[4]]]

非常感谢任何更好的想法或提示

Answer 1

这不像其他答案那样紧凑和节省内存，但希望它应该更容易理解和扩展。

我建议的方法是将您的输入解析为例如列表的列表，每个元素包含一个化合物。我建议 3 个步骤：（1）将数据解析为化合物列表，（2）迭代此化合物列表，删除不需要的化合物，（3）将列表输出回文件。根据文件的大小，可以对数据进行 1 次循环，也可以使用 3 次单独的循环。

# Step (1) Parse the file
compounds = list() # store all compunds
with open('compound.txt', 'r') as f:
    # stores a single compound as a list of rows for a given compound.
    # Note: can be improved to e.g. a dictionary or a custom class
    current_compound = list()
    for line in f:
        if line.strip() == '': # assumes each compound is split by empty line(s)
            print('Empty line')
            # Store previous compound
            if len(current_compound) != 0:
                compounds.append(list(current_compound))

            # prepare for next compound
            current_compound = list()
        else:
            # At this point we could parse this more,
            # e.g. seperate into key/value, but lets just append the whole line with trailing newline
            print('Adding', line.strip())
            current_compound.append(line)

好的，现在让我们检查一下我们的进度

for item in compounds:
    print('\n===Compound===\n', item)

结果

===Compound===
 ['NAME: C11H11NO5; PlaSMA ID-967\n', 'PRECURSORMZ: 238.0712\n', 'PRECURSORTYPE: [M+H]+\n', 'FORMULA: C11H11NO5\n', 'Ontology: Formula predicted\n', 'INCHIKEY:\n', 'SMILES:\n'\
, 'RETENTIONTIME: 1.74\n', 'CCS: -1\n', 'IONMODE: Positive\n', 'COLLISIONENERGY:\n', 'Comment: Annotation level-3; PlaSMA ID-967; ID title-AC_Bulb_Pos-629; Max plant tissue-LE\
_Ripe_Pos\n', 'Num Peaks: 2\n', '192.06602   53\n', '238.0757    31\n']

===Compound===
 ['NAME: Malvidin-3,5-di-O-glucoside; PlaSMA ID-3141\n', 'PRECURSORMZ: 656.19415\n', 'PRECURSORTYPE: [M+H]+\n', 'FORMULA: C29H35O17\n', 'Ontology: Anthocyanidin O-glycosides\n\
', 'INCHIKEY: CILLXFBAACIQNS-UHFFFAOYNA-O\n', 'SMILES: COC1=CC(=CC(OC)=C1O)C1=C(OC2OC(CO)C(O)C(O)C2O)C=C2C(OC3OC(CO)C(O)C(O)C3O)=CC(O)=CC2=[O+]1\n', 'RETENTIONTIME: 2.81\n', '\
CCS: 241.3010517\n', 'IONMODE: Positive\n', 'COLLISIONENERGY:\n', 'Comment: Annotation level-1; PlaSMA ID-3141; ID title-Malvidin-3,5-di-O-glucoside; Max plant tissue-Standard\
 only\n', 'Num Peaks: 0\n']

然后，您可以遍历此化合物列表并删除 Num Peaks 设置为 0 的化合物，然后再写回文件。如果您在这部分也需要帮助，请告诉我。

Answer 2

# Open / read tmp file created with the text you supplied
filedat = open('tmpWrt.txt','r')
filelines = filedat.readlines()

# Open output file object
file_out = open('tmp_out.txt','w')

line_count = 0

# Iterate through all file lines
for line in filelines:
    # If line is beginning of section
    # reset tmp variables
    if line != "\n" and line.split()[0] == "NAME:":
        tmp_lines = []
        flag = 'n'

    tmp_lines.append(line)
    line_count += 1

    # If line is the end of a section and peaks > 0
    # write to file
    if (line == "\n" or line_count == len(filelines)) and flag == 'y':
        #tmp_lines.append("\n")
        for tmp_line in tmp_lines:
            file_out.write(tmp_line)

    # If peaks > 0 set flag to "y"
    if line != "\n" and line.split()[0] == "Num":
            if int(line.split()[2]) != 0:
                flag = "y"

file_out.close()

Answer 3

这是处理文件的一种相当简单的方法。

打开数据文件并遍历其行，将它们存储在列表（缓存）中。如果一行以NAME: 开头，则此行是新记录的开始，如果不为空，则可以打印缓存。

如果该行以Num Peaks: 开头，则检查该值。如果为零，则缓存被清空，导致该记录被遗忘。

仅包含空格的行将被跳过。

with open('data') as f:
    line_cache = []
    for line in f:
        if line.startswith('NAME:'):
            if line_cache:
                print(*line_cache, sep='')
                line_cache = []
        elif line.startswith('Num Peaks:'):
            num_peaks = int(line.partition(': ')[2])
            if num_peaks == 0:
                line_cache = []
                continue

        if line.strip():        # filter empty lines
            line_cache.append(line)

    if line_cache:    # don't forget the last record
        print(*line_cache, sep='', end='')

输出到标准输出。它可以被重定向到你的 shell 环境中的一个文件。如果您想直接写入文件，您可以在开始时打开它并修改print() 语句：

with open('output', 'w') as output, open('data') as f:
    ...

并将print()s 更改为

print(*line_cache, sep='', file=output)