从具有特定模式的文本中获取tsv

Question

我是生物学家，我需要在一个文本文件中获取信息 我有一个纯文本文件，如下所示：

12018411
Comparison of two timed artificial insemination (TAI) protocols for management of first insemination postpartum. 
  TAI|timed artificial insemination|0.999808
Two estrus-synchronization programs were compared and factors influencing their success over a year were evaluated. All cows received a setup injection of PGF2alpha at 39 +/- 3 d postpartum. Fourteen days later they received GnRH, followed in 7 d by a second injection of PGF2alpha. Cows (n = 523) assigned to treatment 1 (modified targeted breeding) were inseminated based on visual signs of estrus at 24, 48, or 72 h after the second PGF2alpha injection. Any cow not observed in estrus was inseminated at 72 h. Cows (n = 440) assigned to treatment 2 received a second GnRH injection 48 h after the second PGF2alpha, and all were inseminated 24 h later. Treatment, season of calving, multiple birth, estrual status at insemination, number of occurrences of estrus before second PGF2alpha, prophylactic use of PGF2alpha, retained fetal membranes, and occurrence of estrus following the setup PGF2alpha influenced success. Conception rate was 31.2% (treatment 1) and 29.1% (treatment 2). A significant interaction occurred between protocol and estrual status at insemination. Cows in estrus at insemination had a 45.8% (treatment 1) or 35.4% (treatment 2) conception rate. The conception rate for cows not expressing estrus at insemination was 19.2% (treatment 1) and 27.7% (treatment 2). Provided good estrous detection exists, modified targeted breeding can be as successful as other timed artificial insemination programs. Nutritional, environmental, and management strategies to reduce postpartum disorders and to minimize the duration of postpartum anestrus are critical if synchronization schemes are used to program first insemination after the voluntary waiting period. 
    
8406022
Deletion of the beta-turn/alpha-helix motif at the exon 2/3 boundary of human c-Myc leads to the loss of its immortalizing function. 
The protein product (c-Myc) of the human c-myc proto-oncogene carries a beta-turn/alpha-helix motif at the exon2/exon3 boundary. The amino acid (aa) sequence and secondary structure of this motif are highly conserved among several nuclearly localized oncogene products, c-Myc, N-Myc, c-Fos, SV40 large T and adenovirus (Ad) Ela. Removal of this region from Ad E1a results in the loss of the transforming properties of the virus without destroying its known transregulatory functions. In order to analyse whether deletion of the above-mentioned region from c-Myc has a similar effect on its transformation activity, we constructed a deletion mutant (c-myc delta) lacking the respective aa at the exon2/exon3 boundary. In contrast to the c-myc wild-type gene product, constitutive expression of c-myc delta does not lead to the immortalization of primary mouse embryo fibroblast cells (MEF cells). This result indicates that c-Myc and Ad El a share a common domain which is involved in the transformation process by both oncogenes. 
  aa|amino acid|0.99818
  Ad|adenovirus|0.96935
  MEF cells|mouse embryo fibroblast cells|0.994648

第一行是id，第二行是标题，第三行是摘要（有时有缩写），最后一行（如果有）是带双空格的缩写、缩写、含义和数字。你可以看到：

  GA|general anesthesia|0.99818

然后在空白处有一行，重新开始：ID、标题、摘要、缩写或ID、标题、缩写、摘要。 我需要将这些数据转换成TSV文件，如下所示：

12018411  TAI  timed artificial insemination
8406022   aa   amino acids
8406022   Ad   adenovirus
...      ...    ...

第一列ID、第二列缩写和该缩写的第三列含义。 我尝试先在数据帧中转换，然后再转换为TSV，但我不知道如何使用我需要的结构获取文本信息。 我也试过使用这个代码：

from collections import namedtuple
import pandas as pd

Item= namedtuple('Item', 'ID')
items = []

with open("identify_abbr-out.txt", "r", encoding='UTF-8') as f:
    lines= f.readlines()
    for line in lines:
        if line== '\n':
            ID= ¿nextline?
        if line.startswith("  "):
            Abbreviation = line
            items.append(Item(ID, Abbreviation))

df = pd.DataFrame.from_records(items, columns=['ID', 'Abbreviation'])

但是我不知道怎么读下一行，没有找到代码，因为有时在语料库和标题之间有空格。p>

我正在使用python 3.8

事先非常感谢

Answer 1

假设test.txt有您的输入数据，我使用简单的文件读取函数来处理数据-

file1 = open('test.txt', 'r') 
Lines = file1.readlines() 
outputlines = []
outputline=""
counter = 0
for l in Lines:
    if l.strip()=="":
        outputline = ""
        counter = 0
    elif counter==0:
        outputline = outputline + l.strip() + "|"
        counter = counter + 1
    elif counter==1:
        counter = counter + 1
    else:
        if len(l.split("|"))==3 and l[0:2]=="  " :
            outputlines.append(outputline + l.strip() +"\n")
        counter = counter + 1
        
file1 = open('myfile.txt', 'w') 
file1.writelines(outputlines) 
file1.close() 

在这里，一行一行地读取文件，保留一个计数器，并在有空行时重置，然后在下一行读取ID。如果有3个字段“|”分隔行，且开头有两个空格，则将使用ID导出该行