I'm reading a file (input.txt) and trying to split the contents with '*' as the delimiter and below is the code I came up with but it assigns all the values to the first list i.e. uno instead of splitting equally. Where I'm doing it wrong. Many thanks for any help.
One more scenario - how to split the contents equally even though we have no idea how many lines are separated with * .In other words, here 3 rows are separated with the delimiter of multiple *s. What if in case of 1000 + rows and 1000 + delimiter of multiple *s.
input.txt
Record: C:/my_files/00_Roll_Tom-values.txt
#RakeBoss-Random as on 12/19/2016
[groups]
met = chk\rel_io_chk, chk\dev_op_io, chk\yuijn7, chk\rft65y
div = chk\ret5cf, chk\plo09i, chk\czo987, chk\lopikm, chk\qzt0kc
rakeonly = chk\r7ynsd, chk\bhtw5h, chk\ko9ibv, chk\plot9q
*******************************************************************************************
Record: C:/my_files/Docker_red-values.txt
#RakeBoss-Check it io.
[groups]
met = chk\rel_io_chk, chk\dev_op_io, chk\yuijn7
div = chk\gzlqvg, chk\k7ygyp, chk\lzg0rp, chk\oli6iv
rakeonly = chk\qzliu0, chk\pl6w6t, chk\bzp0l, chk\dzfbhp
*******************************************************************************************
Record: C:/my_files/456_Milo_123-values.txt
#RakeBoss-Jan 21st Prod
[groups]
met = chk\rel_io_chk, chk\dev_op_io
div = chk\ret5cf, chk\plo09i, chk\cz0o9t, chk\lopikm
rakeonly = chk\ztylcd, chk\hrft9l, chk\zzkkmi
*******************************************************************************************
Code,
file = File.open("C:/rake_check/input.txt", "r")
file.each_line.map do |line|
next if line.chomp! =~ /^$|#/
uno, deus, id = line.split(/\*+/) # also tried line.split("*")
puts second
end
myarray = File.read('input.txt').split(/^\*{91}\n/)
(or, for example, /^\*{5,}\n/ "split on a line of five or more asterisks" if you don't have exactly 91 each time.)
/\*+/ means "one or more asterisks", which has several dangers: you might get an asterisk somewhere inside the record and split on it, and you're not slurping up the newline so each record except the first will start off with one.
Related
I would like to compare two of my log files generated before and after an implementation to see if it has impacted anything. However, the order of the logs I get is not the same all the time. Since, the log file also has multiple indented lines, when I tried to sort, everything is sorted. But, I would like to keep the child intact with the parent. Indented lines are spaces and not tab.
Any help would be greatly appreciated. I am fine with any windows solution or Linux one.
Eg of the file:
#This is a sample code
Parent1 to be verified
Child1 to be verified
Child2 to be verified
Child21 to be verified
Child23 to be verified
Child22 to be verified
Child221 to be verified
Child4 to be verified
Child5 to be verified
Child53 to be verified
Child52 to be verified
Child522 to be verified
Child521 to be verified
Child3 to be verified
I am posting another answer here to sort it hierarchically, using python.
The idea is to attach the parents to the children to make sure that the children under the same parent are sorted together.
See the python script below:
"""Attach parent to children in an indentation-structured text"""
from typing import Tuple, List
import sys
# A unique separator to separate the parent and child in each line
SEPARATOR = '#'
# The indentation
INDENT = ' '
def parse_line(line: str) -> Tuple[int, str]:
"""Parse a line into indentation level and its content
with indentation stripped
Args:
line (str): One of the lines from the input file, with newline ending
Returns:
Tuple[int, str]: The indentation level and the content with
indentation stripped.
Raises:
ValueError: If the line is incorrectly indented.
"""
# strip the leading white spaces
lstripped_line = line.lstrip()
# get the indentation
indent = line[:-len(lstripped_line)]
# Let's check if the indentation is correct
# meaning it should be N * INDENT
n = len(indent) // len(INDENT)
if INDENT * n != indent:
raise ValueError(f"Wrong indentation of line: {line}")
return n, lstripped_line.rstrip('\r\n')
def format_text(txtfile: str) -> List[str]:
"""Format the text file by attaching the parent to it children
Args:
txtfile (str): The text file
Returns:
List[str]: A list of formatted lines
"""
formatted = []
par_indent = par_line = None
with open(txtfile) as ftxt:
for line in ftxt:
# get the indentation level and line without indentation
indent, line_noindent = parse_line(line)
# level 1 parents
if indent == 0:
par_indent = indent
par_line = line_noindent
formatted.append(line_noindent)
# children
elif indent > par_indent:
formatted.append(par_line +
SEPARATOR * (indent - par_indent) +
line_noindent)
par_indent = indent
par_line = par_line + SEPARATOR + line_noindent
# siblings or dedentation
else:
# We just need first `indent` parts of parent line as our prefix
prefix = SEPARATOR.join(par_line.split(SEPARATOR)[:indent])
formatted.append(prefix + SEPARATOR + line_noindent)
par_indent = indent
par_line = prefix + SEPARATOR + line_noindent
return formatted
def sort_and_revert(lines: List[str]):
"""Sort the formatted lines and revert the leading parents
into indentations
Args:
lines (List[str]): list of formatted lines
Prints:
The sorted and reverted lines
"""
sorted_lines = sorted(lines)
for line in sorted_lines:
if SEPARATOR not in line:
print(line)
else:
leading, _, orig_line = line.rpartition(SEPARATOR)
print(INDENT * (leading.count(SEPARATOR) + 1) + orig_line)
def main():
"""Main entry"""
if len(sys.argv) < 2:
print(f"Usage: {sys.argv[0]} <file>")
sys.exit(1)
formatted = format_text(sys.argv[1])
sort_and_revert(formatted)
if __name__ == "__main__":
main()
Let's save it as format.py, and we have a test file, say test.txt:
parent2
child2-1
child2-1-1
child2-2
parent1
child1-2
child1-2-2
child1-2-1
child1-1
Let's test it:
$ python format.py test.txt
parent1
child1-1
child1-2
child1-2-1
child1-2-2
parent2
child2-1
child2-1-1
child2-2
If you wonder how the format_text function formats the text, here is the intermediate results, which also explains why we could make file sorted as we wanted:
parent2
parent2#child2-1
parent2#child2-1#child2-1-1
parent2#child2-2
parent1
parent1#child1-2
parent1#child1-2#child1-2-2
parent1#child1-2#child1-2-1
parent1#child1-1
You may see that each child has its parents attached, all the way along to the root. So that the children under the same parent are sorted together.
Short answer (Linux solution):
sed ':a;N;$!ba;s/\n /#/g' test.txt | sort | sed ':a;N;$!ba;s/#/\n /g'
Test it out:
test.txt
parent2
child2-1
child2-1-1
child2-2
parent1
child1-1
child1-2
child1-2-1
$ sed ':a;N;$!ba;s/\n /#/g' test.txt | sort | sed ':a;N;$!ba;s/#/\n /g'
parent1
child1-1
child1-2
child1-2-1
parent2
child2-1
child2-1-1
child2-2
Explanation:
The idea is to replace the newline followed by an indentation/space with a non newline character, which has to be unique in your file (here I used # for example, if it is not unique in your file, use other characters or even a string), because we need to turn it back the newline and indentation/space later.
About sed command:
:a create a label 'a'
N append the next line to the pattern space
$! if not the last line, ba branch (go to) label 'a'
s substitute, /\n / regex for newline followed by a space
/#/ a unique character to replace the newline and space
if it is not unique in your file, use other characters or even a string
/g global match (as many times as it can)
I have two large files. One of them is an info file(about 270MB and 16,000,000 lines) like this:
1101:10003:17729
1101:10003:19979
1101:10003:23319
1101:10003:24972
1101:10003:2539
1101:10003:28242
1101:10003:28804
The other is a standard FASTQ format(about 27G and 280,000,000 lines) like this:
#ST-E00126:65:H3VJ2CCXX:7:1101:1416:1801 1:N:0:5
NTGCCTGACCGTACCGAGGCTAACCCTAATGAGCTTAATCAAGATGATGCTCGTTATGG
+
AAAFFKKKKKKKKKFKKKKKKKFKKKKAFKKKKKAF7AAFFKFAAFFFKKF7FF<FKK
#ST-E00126:65:H3VJ2CCXX:7:1101:10003:75641:N:0:5
TAAGATAGATAGCCGAGGCTAACCCTAATGAGCTTAATCAAGATGATGCTCGTTATGG
+
AAAFFKKKKKKKKKFKKKKKKKFKKKKAFKKKKKAF7AAFFKFAAFFFKKF7FF<FKK
The FASTQ file uses four lines per sequence. Line 1 begins with a '#' character and is followed by a sequence identifie. For each sequence,this part of the Line 1 is unique.
1101:1416:1801 and 1101:10003:75641
And I want to grab the Line 1 and the next three lines from the FASTQ file according to the info file. Here is my code:
import gzip
import re
count = 0
with open('info_path') as info, open('grab_path','w') as grab:
for i in info:
sample = i.strip()
with gzip.open('fq_path') as fq:
for j in fq:
count += 1
if count%4 == 1:
line = j.strip()
m = re.search(sample,j)
if m != None:
grab.writelines(line+'\n'+fq.next()+fq.next()+fq.next())
count = 0
break
And it works, but because both of these two files have millions of lines, it's inefficient(running one day only get 20,000 lines).
UPDATE at July 6th:
I find that the info file can be read into the memory(thank #tobias_k for reminding me), so I creat a dictionary that the keys are info lines and the values are all 0. After that, I read the FASTQ file every 4 line, use the identifier part as the key,if the value is 0 then return the 4 lines. Here is my code:
import gzip
dic = {}
with open('info_path') as info:
for i in info:
sample = i.strip()
dic[sample] = 0
with gzip.open('fq_path') as fq, open('grap_path',"w") as grab:
for j in fq:
if j[:10] == '#ST-E00126':
line = j.split(':')
match = line[4] +':'+line[5]+':'+line[6][:-2]
if dic.get(match) == 0:
grab.writelines(j+fq.next()+fq.next()+fq.next())
This way is much faster, it takes 20mins to get all the matched lines(about 64,000,000 lines). And I have thought about sorting the FASTQ file first by external sort. Splitting the file that can be read into the memory is ok, my trouble is how to keep the next three lines following the indentifier line while sorting. The Google's answer is to linear these four lines first, but it will take 40mins to do so.
Anyway thanks for your help.
You can sort both files by the identifier (the 1101:1416:1801) part. Even if files do not fit into memory, you can use external sorting.
After this, you can apply a simple merge-like strategy: read both files together and do the matching in the meantime. Something like this (pseudocode):
entry1 = readFromFile1()
entry2 = readFromFile2()
while (none of the files ended)
if (entry1.id == entry2.id)
record match
else if (entry1.id < entry2.id)
entry1 = readFromFile1()
else
entry2 = readFromFile2()
This way entry1.id and entry2.id are always close to each other and you will not miss any matches. At the same time, this approach requires iterating over each file once.
I'm moving my bookmarks from kippt.com to pinboard.in.
I exported my bookmarks from Kippt and for some reason, they were storing tags (preceded by #) and description within the same field. Pinboard keeps tags and description separated.
This is what a Kippt bookmark looks like after export:
<DT>This is a title
<DD>#tag1 #tag2 This is a description
This is what it should look like before importing into Pinboard:
<DT>This is a title
<DD>This is a description
So basically, I need to replace #tag1 #tag2 by TAGS="tag1,tag2" and move it on the first line within <A>.
I've been reading about moving chunks of data here: sed or awk to move one chunk of text betwen first pattern pair into second pair?
I haven't been to come up with a good recipe so far. Any insight?
Edit:
Here's an actual example of what the input file looks like (3 entries out of 3500):
<DT>Phabricator
<DD>#bug #tracking
<DT>The hidden commands for diagnosing and improving your Netflix streaming quality – Quartz
<DT>Icelandic Farm Holidays | Local experts in Iceland vacations
<DD>#iceland #tour #car #drive #self Self-driving tour of Iceland
This might not be the most beautiful solution, but since it seems to be a one-time-thing it should be sufficient.
import re
dt = re.compile('^<DT>')
dd = re.compile('^<DD>')
with open('bookmarks.xml', 'r') as f:
for line in f:
if re.match(dt, line):
current_dt = line.strip()
elif re.match(dd, line):
current_dd = line
tags = [w for w in line[4:].split(' ') if w.startswith('#')]
current_dt = re.sub('(<A[^>]+)>', '\\1 TAGS="' + ','.join([t[1:] for t in tags]) + '">', current_dt)
for t in tags:
current_dd = current_dd.replace(t + ' ', '')
if current_dd.strip() == '<DD>':
current_dd = ""
else:
print current_dt
print current_dd
current_dt = ""
current_dd = ""
print current_dt
print current_dd
If some parts of the code are not clear, just tell me. You can of course use python to write the lines to a file instead of printing them, or even modify the original file.
Edit: Added if-clause so that empty <DD> lines won't show up in the result.
script.awk
BEGIN{FS="#"}
/^<DT>/{
if(d==1) print "<DT>"s # for printing lines with no tags
s=substr($0,5);tags="" # Copying the line after "<DT>". You'll know why
d=1
}
/^<DD>/{
d=0
m=match(s,/>/) # Find the end of the HREF descritor first match of ">"
for(i=2;i<=NF;i++){sub(/ $/,"",$i);tags=tags","$i} # Concatenate tags
td=match(tags,/ /) # Parse for tag description (marked by a preceding space).
if(td==0){ # No description exists
tags=substr(tags,2)
tagdes=""
}
else{ # Description exists
tagdes=substr(tags,td)
tags=substr(tags,2,td-2)
}
print "<DT>" substr(s,1,m-1) ", TAGS=\"" tags "\"" substr(s,m)
print "<DD>" tagdes
}
awk -f script.awk kippt > pinboard
INPUT
<DT>Phabricator
<DD>#bug #tracking
<DT>The hidden commands for diagnosing and improving your Netflix streaming quality – Quartz
<DT>Icelandic Farm Holidays | Local experts in Iceland vacations
<DD>#iceland #tour #car #drive #self Self-driving tour of Iceland
OUTPUT:
<DT>Phabricator
<DD>
<DT>The hidden commands for diagnosing and improving your Netflix streaming quality – Quartz
<DT>Icelandic Farm Holidays | Local experts in Iceland vacations
<DD> Self-driving tour of Iceland
I am trying to extract information from a large file and cannot figure out how to extract strings from file lines only when a previous line in the same record within the file has been matched by regex. An example of one record in the file is as follows:
*NEW RECORD
RECTYPE = D
MH = Informed Consent
AQ = ES HI LJ PX SN ST
ENTRY = Consent, Informed
MN = N03.706.437.650.312
MN = N03.706.535.489
FX = Disclosure
FX = Mental Competency
FX = Therapeutic Misconception
FX = Treatment Refusal
ST = T058
ST = T078
AN = competency to consent: coordinate IM with MENTAL COMPETENCY (IM)
PI = Jurisprudence (1966-1970)
PI = Physician-Patient Relations (1966-1970)
MS = Voluntary authorization, by a patient or research subject, etc,...
This file contains over 20,000 records like this example. I want to identify a small percent of those records using the "MH" field. In this example, I want to find "Informed Consent", and then use regex to extract the information in the FX, AN, and MS fields only within that record. So far, I have opened the file, accessed the hash that the MH terms are stored in, and been able to extract those terms from the records in the file. I also have a functioning regex that identifies the content in the "FX" field.
File.open('mesh_descriptor.bin').each do |file_line|
file_line = file_line.chomp
# read each key of candidate_descriptor_keys
candidate_descriptor_keys.each do |cand_term|
if file_line =~ /^MH\s=\s(#{cand_term})$/
mesh_header = $1
puts "MH from Mesh Descriptor file is: #{mesh_header}"
if file_line =~ /^FX\s=\s(.*)$/
see_also = $1
puts " See_Also from Descriptor file is: #{see_also}"
end
end
end
end
The hash contains the following MH (keys):
candidate_descriptor_keys = ["Body Weight", "Obesity", "Thinness", "Fetal Weight", "Overweight"]
I had success extracting "FX" when I put the statement outside of the "if" statement to extract "MH", but all of the "FX" from the whole file were retrieved - not what I need. I thought putting the "if" statement for "FX" within the previous "if" statement would restrict the results to only those found when the first statement is true, but I am getting no results (also no errors) with this strategy. What I would like as a result is:
> Informed Consent
> Disclosure
> Mental Competency
> Therapeutic Misconception
> Treatment Refusal
as well as the strings within the "AN" and "MS" fields for only those records matching "MH". Any suggestions would be helpful!
I think this may be what you are looking for, but if not, let me know and I will change it. Look especially at the very end to see if that is the sort of output (for input having two records, both with a "MH" field) you want. I will also add a "explanation" section at the end once I have understood your question correctly.
I have assumed that each record begins
*NEW_RECORD
and you wish to identify all lines beginning "MH" whose field is one of the elements of:
candidate_descriptor_keys =
["Body Weight", "Obesity", "Thinness", "Informed Consent"]
and for each match, you would like to print the contents of the lines for the same record that begin with "FX", "AN" and "MS".
Code
NEW_RECORD_MARKER = "*NEW RECORD"
def getem(fname, candidate_descriptor_keys)
line = 0
found_mh = false
File.open(fname).each do |file_line|
file_line = file_line.strip
case
when file_line == NEW_RECORD_MARKER
puts # space between records
found_mh = false
when found_mh == false
candidate_descriptor_keys.each do |cand_term|
if file_line =~ /^MH\s=\s(#{cand_term})$/
found_mh = true
puts "MH from line #{line} of file is: #{cand_term}"
break
end
end
when found_mh
["FX", "AN", "MS"].each do |des|
if file_line =~ /^#{des}\s=\s(.*)$/
see_also = $1
puts " Line #{line} of file is: #{des}: #{see_also}"
end
end
end
line += 1
end
end
Example
Let's begin be creating a file, starging with a "here document that contains two records":
records =<<_
*NEW RECORD
RECTYPE = D
MH = Informed Consent
AQ = ES HI LJ PX SN ST
ENTRY = Consent, Informed
MN = N03.706.437.650.312
MN = N03.706.535.489
FX = Disclosure
FX = Mental Competency
FX = Therapeutic Misconception
FX = Treatment Refusal
ST = T058
ST = T078
AN = competency to consent
PI = Jurisprudence (1966-1970)
PI = Physician-Patient Relations (1966-1970)
MS = Voluntary authorization
*NEW RECORD
MH = Obesity
AQ = ES HI LJ PX SN ST
ENTRY = Obesity
MN = N03.706.437.650.312
MN = N03.706.535.489
FX = 1st FX
FX = 2nd FX
AN = Only AN
PI = Jurisprudence (1966-1970)
PI = Physician-Patient Relations (1966-1970)
MS = Only MS
_
If you puts records you will see it is just a string. (You'll see that I shortened two of them.) Now write it to a file:
File.write('mesh_descriptor', records)
If you wish to confirm the file contents, you could do this:
puts File.read('mesh_descriptor')
We also need to define define the array candidate_descriptor_keys:
candidate_descriptor_keys =
["Body Weight", "Obesity", "Thinness", "Informed Consent"]
We can now execute the method getem:
getem('mesh_descriptor', candidate_descriptor_keys)
MH from line 2 of file is: Informed Consent
Line 7 of file is: FX: Disclosure
Line 8 of file is: FX: Mental Competency
Line 9 of file is: FX: Therapeutic Misconception
Line 10 of file is: FX: Treatment Refusal
Line 13 of file is: AN: competency to consent
Line 16 of file is: MS: Voluntary authorization
MH from line 18 of file is: Obesity
Line 23 of file is: FX: 1st FX
Line 24 of file is: FX: 2nd FX
Line 25 of file is: AN: Only AN
Line 28 of file is: MS: Only MS
Let's say I have an array of Twitter account names:
string = %w[example1 example2 example3 example4 example5 example6 example7 example8 example9 example10 example11 example12 example13 example14 example15 example16 example17 example18 example19 example20]
And a prepend and append variable:
prepend = 'Check out these cool people: '
append = ' #FollowFriday'
How can I turn this into an array of as few strings as possible each with a maximum length of 140 characters, starting with the prepend text, ending with the append text, and in between the Twitter account names all starting with an #-sign and separated with a space. Like this:
tweets = ['Check out these cool people: #example1 #example2 #example3 #example4 #example5 #example6 #example7 #example8 #example9 #FollowFriday', 'Check out these cool people: #example10 #example11 #example12 #example13 #example14 #example15 #example16 #example17 #FollowFriday', 'Check out these cool people: #example18 #example19 #example20 #FollowFriday']
(The order of the accounts isn't important so theoretically you could try and find the best order to make the most use of the available space, but that's not required.)
Any suggestions? I'm thinking I should use the scan method, but haven't figured out the right way yet.
It's pretty easy using a bunch of loops, but I'm guessing that won't be necessary when using the right Ruby methods. Here's what I came up with so far:
# Create one long string of #usernames separated by a space
tmp = twitter_accounts.map!{|a| a.insert(0, '#')}.join(' ')
# alternative: tmp = '#' + twitter_accounts.join(' #')
# Number of characters left for mentioning the Twitter accounts
length = 140 - (prepend + append).length
# This method would split a string into multiple strings
# each with a maximum length of 'length' and it will only split on empty spaces (' ')
# ideally strip that space as well (although .map(&:strip) could be use too)
tweets = tmp.some_method(' ', length)
# Prepend and append
tweets.map!{|t| prepend + t + append}
P.S.
If anyone has a suggestion for a better title let me know. I had a difficult time summarizing my question.
The String rindex method has an optional parameter where you can specify where to start searching backwards in a string:
arr = %w[example1 example2 example3 example4 example5 example6 example7 example8 example9 example10 example11 example12 example13 example14 example15 example16 example17 example18 example19 example20]
str = arr.map{|name|"##{name}"}.join(' ')
prepend = 'Check out these cool people: '
append = ' #FollowFriday'
max_chars = 140 - prepend.size - append.size
until str.size <= max_chars do
p str.slice!(0, str.rindex(" ", max_chars))
str.lstrip! #get rid of the leading space
end
p str unless str.empty?
I'd make use of reduce for this:
string = %w[example1 example2 example3 example4 example5 example6 example7 example8 example9 example10 example11 example12 example13 example14 example15 example16 example17 example18 example19 example20]
prepend = 'Check out these cool people:'
append = '#FollowFriday'
# Extra -1 is for the space before `append`
max_content_length = 140 - prepend.length - append.length - 1
content_strings = string.reduce([""]) { |result, target|
result.push("") if result[-1].length + target.length + 2 > max_content_length
result[-1] += " ##{target}"
result
}
tweets = content_strings.map { |s| "#{prepend}#{s} #{append}" }
Which would yield:
"Check out these cool people: #example1 #example2 #example3 #example4 #example5 #example6 #example7 #example8 #example9 #FollowFriday"
"Check out these cool people: #example10 #example11 #example12 #example13 #example14 #example15 #example16 #example17 #FollowFriday"
"Check out these cool people: #example18 #example19 #example20 #FollowFriday"