.txt 文件在 python 中处理 ML 数据

Question

我有 .txt 文件，其中包含 ML 的数据。每个文件有 19 个条目，这些条目在文本文件中一次又一次地重复。我需要在 python 中以这样的方式获取这些文件，即每个条目都成为一列，并且它们相应的数值被放置在该列中。 我需要对 txt 文件中设置的所有 19 个条目执行此操作（如下面的“1”）。或仅包含数字数据的文本文件也可以（如下面的“2”）。如果您可以在 python 中为此编写一个很棒的代码，否则请描述一些详细信息如何做到这一点

我想要的输出：

1click to view image 或者 2click to view image

文本文件：

Frame.id：263126，时间戳：697287019071，Hand_number：1 ,hand_Id_type: 238右手 ,hand_finger's_number:2,手方向: (-0.142081, 0.865413, -0.480493) ,手掌位置: (35.2841, 284.522, 330.828) ,手掌法线: (-0.686854, -0.435733, -0.58) ,finger_type(tipposition) TYPE_THUMB(-36.7239, 301.602, 330.845) ,finger_type(tipposition) TYPE_INDEX(-14.9321, 347.039, 280.375) ,finger_type(tipposition) TYPE_MIDDLE(5.5661, 258.191, 321.318) ,finger_type(tipposition) TYPE_RING(20.0886, 251.219, 320.136) ,finger_type(tipposition) TYPE_PINKY(27.5919, 259.584, 310.508) Frame.id：263127，时间戳：697287037765，Hand_number：1 ,hand_Id_type: 238右手 ,hand_finger's_number:2,手方向: (-0.167599, 0.827034, -0.536587) ,手掌位置: (31.7441, 283.449, 322.619) ,手掌法线: (-0.776892, -0.445873, -0.44) ,finger_type(tipposition) TYPE_THUMB(-38.8083, 304.444, 330.446) ,finger_type(tipposition) TYPE_INDEX(-22.0532, 344.008, 273.496) ,finger_type(tipposition) TYPE_MIDDLE(-0.161068, 258.09, 319.489) ,finger_type(tipposition) TYPE_RING(13.8233, 250.236, 317.138) ,finger_type(tipposition) TYPE_PINKY(20.1892, 257.352, 305.739) Frame.id：263128，时间戳：697287057570，Hand_number：1 ,hand_Id_type: 238右手,hand_finger's_number:2,手方向: (-0.179139, 0.817551, -0.547284) ,手掌位置: (30.8754, 280.871, 315.444) ,手掌正常: (-0.750039, -0.473481, -0.473481, -0.461 ,finger_type(tipposition) TYPE_THUMB(-40.4781, 299.689, 321.24) ,finger_type(tipposition) TYPE_INDEX(-23.5209, 339.286, 264.435) ,finger_type(tipposition) TYPE_MIDDLE(-0.157164, 254.483, 311.627) ,finger_type(tipposition) TYPE_RING(14.1716, 247.067, 309.742) ,finger_type(tipposition) TYPE_PINKY(20.45, 254.358, 298.274) Frame.id：263129，时间戳：697287076710，Hand_number：1 ,hand_Id_type: 238右手,hand_finger's_number:2,手方向: (-0.191306, 0.830611, -0.522961) ,掌位: (28.6877, 277.055, 299.705) ,掌正常: (-0.739979, -0.472093, -0.493) ,finger_type(tipposition) TYPE_THUMB(-42.9838, 294.545, 305.915) ,finger_type(tipposition) TYPE_INDEX(-26.5976, 335.972, 250.27) ,finger_type(tipposition) TYPE_MIDDLE(-1.90294, 250.26, 294.934) ,finger_type(tipposition) TYPE_RING(12.5955, 243.157, 292.909) ,finger_type(tipposition) TYPE_PINKY(18.5831, 250.968, 281.465)

文本文件链接：https://drive.google.com/file/d/1X1NdQNYlQWuNpmzGL6Wwi_SqFjIXnRbb/view?usp=sharing

Answer 1

不是解析文件的最复杂的方法，但似乎可行。

data = []

text = []
with open('data_file_1.txt') as f:
    text = f.readlines()

text = ''.join(text) # Join all lines together, just in case
text = text.strip('\n').strip('\t') # Make whole text one line, just in case
rows = text.split('Frame.') # Separate rows of data

for row in rows[1:]: # First split is empty, start at second
    row_data = {}
    expected = 0
    store = []
    for item in row.split(','): # Split data at comma. Note: tuples will be split also
        if expected: # If in last item was a start of tuple, add rest values here
            expected -= 1
            store[1].append(float(item.strip().strip(')')))
            if not expected: # when all values are in place, save whole item to row_data
                row_data[store[0]] = store[1]
                store = []

        # In single value cases, parse and save them
        elif any(x in item for x in ['id', 'Timestamp', 'Hand_number', 'hand_Id_type', "hand_finger's_number"]):
            name, value = item.split(':')
            try:
                value = int(value)
            except ValueError:
                value = value.strip()
            row_data[name.strip()] = value

        # In cases, where there are tuple of 3 floats, parse and store for save
        elif any(x in item for x in ['Finger_type(tipposition)', 'Palm', 'direction']):
            name, value = item.strip('Finger_type(tipposition) ').split('(')
            store = [name.strip().strip(':'), [float(value)]]
            expected = 2

    data.append(row_data)

for row in data: # Final data in list of dict's
    print(row)

# Export your data in format like: .csv, JSON, XML, what is suitable for use in your application

像这样打印行

{'id': 263126, 'Timestamp': 697287019071, 'Hand_number': 1, 'hand_Id_type': '238right hand', "hand_finger's_number": 2, 'hand direction': [-0.142081, 0.865413, -0.480493], 'Palm position': [35.2841, 284.522, 330.828], 'Palm normal': [-0.686854, -0.435733, -0.581694], 'TYPE_THUMB': [-36.7239, 301.602, 330.845], 'TYPE_INDEX': [-14.9321, 347.039, 280.375], 'TYPE_MIDDLE': [5.5661, 258.191, 321.318], 'TYPE_RING': [20.0886, 251.219, 320.136], 'TYPE_PINKY': [27.5919, 259.584, 310.508]}

Answer 2

我做了一些简单的争论并将其制成 CSV。希望这对你有用：

lines = []
with open("data.txt") as f:
    lines=f.readlines()

# Columns in file    
header = "Frame.id,Timestamp,Hand_number,hand_Id_type,hand_finger's_number,hand direction,\
Palm position,Palm normal,Finger_type(tipposition) TYPE_THUMB,\
Finger_type(tipposition) TYPE_INDEX,Finger_type(tipposition) TYPE_MIDDLE,\
Finger_type(tipposition) TYPE_RING,Finger_type(tipposition) TYPE_PINKY\n"
out_lines = [header] # to store rows in output file
tmp=[] # empty list to store one row
for line in lines[1:]:
    tmp.append(line.strip()) # strip and append
    # if line has last column in row, add it to tmp, make the line
    # add the line to list of rows and clear tmp
    if line.strip().startswith(",Finger_type(tipposition) TYPE_PINKY"):
        new_line = ''.join(tmp)+'\n'
        for column in header.split(','):
            # replace header keywords and :
            new_line = new_line.replace(column.strip(), '').replace(':','')
        out_lines.append(new_line)
        tmp=[]
        
print(''.join(out_lines))

# Uncomment the below block to write the csv file     
# with open("output.csv","w+") as f:
#     f.writelines(out_lines)