Convert tab separated text to non-trivial xml. (pwsafe -- KeePassx)

Question

I'd like to take the output of `pwsafe --exportdb > database.txt` and convert it to a KeePassX XML friendly format (feature request in pwsafe).
I found flat file converter but the syntax is beyond me with this example.  Solutions are welcomed.
More details
Here is the pwsafe --> KeePassX XML translations.  The pwsafe export is simply a txt file with 6 fields (the first field can be ignored):
uuid= doesn't translate
group= group>title
name= entry>title
login= entry>username
passwd= entry>password
notes= entry>comment
Example txt file for conversion (exported from pwsafe):
# passwordsafe version 2.0 database"
uuid group name login passwd notes
"123d9-daf-df-3423423" "retail" "amazon" "myamazonuser" "sjfJ849" "superfluous comment"
"4599d934-dsfs-324" "retail" "netflix" "netflixuser" "dj3W$#" ""
"4kdfkd-434-jj" "email" "gmail" "mygmail" "dfkpass" ""
Example xml in keepassx xml:
<!DOCTYPE KEEPASSX_DATABASE>
<database>
<group>
<title>Internet</title>
<entry>
<title>github</title>
<username>githubusername</username>
<password>githubpassword</password>
<comment>optional comment</comment>
</entry>
</group>
<group>
<title>retail</title>
<entry>
<title>amazon</title>
<username>username</username>
<password>myamazonpw</password>
</entry>
</group>
<group>
<title>retail</title>
<entry>
<title>netflix</title>
<username>username</username>
<password>mynfxpw</password>
</entry>
</group>
</database>
Last edited by graysky (2015-06-14 18:27:17)

Answer 1

It seems that ffe does not work quite well with tab/space separated list (at least I was not able to get it working). However you can use sed to replace tabs with commas:
sed -e "s/\t/,/g" input.txt | ffe -c pwsafe.rc -s example.txt -l 2>/dev/null
and this is pwsafe.rc
structure pwsafe {
type separated , *
output group
quoted
header first
record line {
field uuid * * noprint
field title
field title * * entry
field username * * entry
field passwordd * * entry
field comment * * entry
output noprint {
data ""
no-data-print no
output entry {
data " <%n>%t</%n>\n"
no-data-print no
output group {
file_header "<!DOCTYPE KEEPASSX_DATABASE>\n <database>\n"
record_header " <group>\n"
data " <%n>%t</%n>\n <entry>\n"
record_trailer " <entry>\n </group>\n"
file_trailer "</database>\n"
no-data-print no
If you are not bound to ffe, maybe it is easier to use a python script which is simpler to modify and it is more flexible. Something like the following should work:
#!/usr/bin/python
import sys
from xml.dom import minidom
class Converter(object):
def __init__(self, filename):
self.url = filename
def convert(self):
inp_f = open(self.url, 'r')
data = inp_f.readlines()
inp_f.close()
# xml document model
doc = minidom.Document()
root = doc.createElement('database')
doc.appendChild(root)
for line in data:
if '"' not in line:
continue
fields = line.split('\t')
if len(fields) < 6:
continue
# uuid = fields[0].strip('"') # unused
group = fields[1].strip('" ')
name = fields[2].strip('" ')
login = fields[3].strip('" ')
passwd = fields[4].strip('" ')
notes = fields[5].strip('" \n')
group_node = doc.createElement('group')
root.appendChild(group_node)
# <group>
group_title_node = doc.createElement('title')
group_node.appendChild(group_title_node)
group_title_node.appendChild(doc.createTextNode(group))
# one <entry> per <group>
entry_node = doc.createElement('entry')
group_node.appendChild(entry_node)
# <entry> -> <title>
entry_title_node = doc.createElement('title')
entry_title_node.appendChild(doc.createTextNode(name))
entry_node.appendChild(entry_title_node)
# <entry> -> <username>
entry_uname_node = doc.createElement('username')
entry_uname_node.appendChild(doc.createTextNode(login))
entry_node.appendChild(entry_uname_node)
# <entry> -> <password>
entry_passwd_node = doc.createElement('password')
entry_passwd_node.appendChild(doc.createTextNode(passwd))
entry_node.appendChild(entry_passwd_node)
# <entry> -> <comments>
entry_comment_node = doc.createElement('comment')
entry_comment_node.appendChild(doc.createTextNode(notes))
entry_node.appendChild(entry_comment_node)
print('<!DOCTYPE KEEPASSX_DATABASE>')
print(root.toprettyxml(' ', '\n'))
if __name__ == "__main__":
try:
ifname = sys.argv[1]
except IndexError:
print("Input file name required")
sys.exit(1)
cc = Converter(ifname)
cc.convert()
NOTE: I'm assuming from the example you posted that every <group> node contains only one <entry> child node.
--edit[0]: corrected a typo in the python code (just realized that it is <comment> instead of <comments ) and made the parsing function more reliable
Last edited by mauritiusdadd (2015-06-15 09:29:01)