#!/usr/bin/env python

import ftplib
import time
import datetime

def getValue(tokens,kw) :
    """
    Return a value for a key/value pair, where
    = is the separator.
    """
    i = 0
    L = len(tokens)
    FOUND = False
    retval = ""
    while i < L  and not FOUND :
        t = tokens[i].split('=')
        if t[0] == kw :
            FOUND = True
            retval = t[1]
        i += 1
    return retval

#These servers support MLSD
FTPINFO = {"ftp.ncbi.nih.gov" : "/blast/db" , "ftp.hgc.jp" : "pub/mirror/ncbi/blast/db",
     "ftp.ebi.ac.uk" : "pub/blast/db"}

#These servers don't support MLSD
OTHERFTPINFO = {"bio-mirror.net" : "/blast/db" , "mirrors.vbi.vt.edu" : "mirrors/ftp.ncbi.nih.gov/blast/db"}

FTPSITE = 'ftp.ncbi.nih.gov'
#FTPSITE = 'ftp.ebi.ac.uk'
#FTPSITE = 'ftp.hgc.jp'
FTPDIR = FTPINFO[FTPSITE]

ftp = ftplib.FTP(FTPSITE)
ftp.login()
ftp.cwd(FTPDIR)

SupportsMLSD = True
dirlines = []
try :
    ftp.retrlines('MLSD',dirlines.append)
    # Python 3 - use ftp.mlsd
except ftplib.all_errors as e :
    SupportsMLSD = False 
    print(str(e).split(None,1))
    print('Site doesn\'t support MLSD')
    ftp.retrlines('LIST',dirlines.append)


if SupportsMLSD :
    for line in dirlines:
        tokens = line.split(';')
        numtokens = len(tokens)
        Name = tokens[numtokens-1].lstrip()
        Mtime = getValue(tokens,'modify')
        Size = getValue(tokens,'size')
        print(Name + ' ' + Size + ' ' + str(Mtime))
else:
    now = datetime.datetime.now()
    CurrentYear = now.year
    for line in dirlines:

        """
        Parsing directory lines is hazardous!
        Directory listings from a number of FTP sites seem to
        all be consistent with tokenizing into nine non-blank fields:
        Field   Content
        -----------------------
        0       permissions
        1       ?
        2       owner
        3       group
        4       size
        5       Month
        6       Day
        7       Time or Year
        8       Name
        
        Field 8, is a special case, because a name might include blanks.
        For the purposes of downloading NCBI files, we don't need to worry
        about blanks in names.
        """

        # Parse directory lines into non-blank fields
        # When None is used as the separator, multiple seperators are
        # parsed as a single seperator. Neat!
        tokens = line.split(None)
        #print(line)
        L = len(tokens)
        if L == 9 :
            # Name and Size are easy
            Name = tokens[8]
            Size = tokens[4]

            # Modification time takes a bit more work.
            Month = tokens[5]
            Day = tokens[6]
            if ':' in tokens[7] : 
                Year = str(CurrentYear)
                TimeHM = tokens[7]
            else :
                Year = tokens[7]
                TimeHM = "00:00"
            Mtime = Day + ' ' +  Month + ' ' + Year + ' ' + TimeHM 
            print(Name + ' ' + Size + ' ' + Mtime)       
            Mtime = time.strptime(Mtime, "%d %b %Y %H:%M")
            print(Name + ' ' + Size + ' ' + str(Mtime))

ftp.quit()

""" 
For FTP servers that support MLSD, it's easy
to get the metadata.

For those that don't, ftp has commands that
can be used to get size and mtime:


ftp> modtime robots.txt
robots.txt	02/17/2015 17:01:36 GMT

ftp> size robots.txt
213 26

This is a pain in the ass to work with, 
but it's better than nothing.

"""