Eduroam and IRC

September 23rd, 2013 § 0 comments § permalink

SSH tunnelling out of eduroam network.

The six words version of a post looks promising and challenging.

So… I’m sitting somewhere in the university campus, trying to access IRC, connected through an eduroam node… and I can’t.

Turns out you can’t1 do much more other than web or mail traffic2. Bummer.

So the first thing popping to my mind is to use a VPN, but those services are sort of expensive when on a student budget.

Idea #2 is to SSH tunnel but, as is, that’s sort of a bummer3. So why not use sshuttle? I mean, it’s still tunnelling but now it’s transparent (all your network will tunnel through it) and you don’t need to change any of your configuration. All you need is still an external machine where you can ssh into and from where you can connect to you desired service.

Just run ./shuttle username@domain 0/0 and you’ll be fine!

PS. This post was hanging here for so long, so I’m just posting it mostly as I found it, have fun!

  1. I don’t know if this is general eduroam policy or just my university policy. []
  2. http://www.wifi.polimi.it/en/wi-fi/eduroam-network/ []
  3. You need to reconfigure your client everytime you switch from tunnelling vs non tunnelling to enable/disable the proxy for your client. []

What happened?

December 6th, 2010 § 0 comments § permalink

Dataloss makes me bored, not anymore.

That would be the six words version of the story. I had problems with my host and didn’t have the latest backup of the blog, but I discovered it only after putting everything back online on another host. Other than that there was to fix all the 3rd domain names, upload images and build again a theme: it pissed me off for a while.

I left the blog online as a zombie, there are few posts left and I plan to leave them there. I’m starting to polish the blog starting with the theme and going to post some interesting stuff in the near future, this could be a fresh1 start.

Let’s see if any old feed reader pops out to say hello!

  1. Let’s say almost fresh. Did you see? I installed WP-Footnotes too. []

A geeky toy, for a geeky boy.

March 3rd, 2009 § 1 comment § permalink

I’m sure all of you know what the rubik’s cube is, if you don’t… I mean, really? man what was you thinking until now? Well If you REALLY don’t know it then you should at least google for it. Wikipedia will work as well.

rubik's cube

I always thought it was a logic puzzle, and after some unsuccessful tries I decided to cheat, googled for “rubik’s cube solution” and I found out a beautiful site, which explains everything you need to solve the cube in a very simple and fast way, it is of course the Lars Petrus method.

Once learnt a solution method there is no more “logic” in it. Where is the fun part? What’s the challenge?
I was terribly wrong. It is not a logic-puzzle, it is a memory-puzzle, and a hand-dexterity-puzzle.

You need to remember the moves of your solving methods, they are almost always the same (in the petrus method there are 4 basic moves), but the most challenging part is how fast can you solve the cube?

Hey “how-fast” sounds damn challenging, how can you measure that? You’ll need a stopwatch. If you are at least half geeky as I am you’d be already on google at the moment and starting to type something like “online stopwatch” or “online timer”.

There are tons of them, but all have a very annoying interface that make you use the mouse. Given that you are a very good speedcuber you don’t want to loose 0.3s to start to solve and an other 0.3s to stop the timer, do you?

Well I wouldn’t care as I need 1’30”+ to solve the cube, but I found fun to code a stopwatch designed to work with Miss Spacebar instead of your loveable mice.

nigma timer

the simple nigma timer is written in simple html/css/js and as a very simple interface, I hope you’ll enjoy it.

return ”

How to convert Microsoft Word .doc files to PDF from command line

January 14th, 2009 § 1 comment § permalink

I know lot of people need it, Google is full of requests by hundred, maybe thousands of users asking for a doc2pdf converter or this kind of thing. I need it too. It is useful to have all files in pdf format (and maybe all merged in one file only) and if you have a lot of files to convert by hand, believe me, you’re not going to have a nice day.

The easy way

It is pretty easy:

$ abiword --to=pdf filename.doc

I don’t think there is so much to explain here. It converts filename.doc to filename.pdf and saves it in the current directory. It was too easy. Why should you need an hard way? I don’t know, I’m sure I need one. Unfortunately abiword’s Microsoft doc file support is not so good, in fact it lacks of the math and image/clipart features. I’m not sure if this affects all versions of abiword but it is sure for the one that comes with ubuntu (actually it doesn’t come with it, you’ve to apt-get install it).

Anyway I really need to see plots and formulas. What you said? OpenOffice supports them. Check it out. Yes I know that, OpenOffice can read almost always plots and images in doc files. Bad luck seems to be here again, OpenOffice lacks of the same command line interface abiword has, so the only way is to open doc files one by one and click on the Export as PDF button. It is very frustrating. So, here is the hard way.

The hard way

Short version (for whom doesn’t like read me be but want to read so much): check the Python-UNO site.

Long version. You need to know what Python-UNO is

The Python-UNO bridge allows to

  • use the standard OpenOffice.org API from the well known python scripting language.
  • to develop UNO components in python, thus python UNO components may be run within the OpenOffice.org process and can be called from Java, C++ or the built in StarBasic scripting language.
  • create and invoke scripts with the office scripting framework (OOo 2.0 and later).

You can find the most current version of this document from http://udk.openoffice.org/python/python-bridge.html

Oh no! I’ll have to download this Python-UNO, read manuals to learn how to use those API and who knows if it’ll work…… No. Just don’t panic. I’m going to tell you something that will make this a not-so-hard way. The first thing is that if you have installed OpenOffice you’re at 50% of the work, in fact Pyhton-UNO comes with OpenOffice since version 1.1.

  • Pyhton-UNO comes with OpenOffice since version 1.1. You don’t have to download and install anything
  • Pyhton-UNO’s guys are so cool that in their code examples there is all of what we need.

From the examples page you can download the ooextract.py script. It has a very simple usage, we need to use it in this way:

$ openoffice -invisible "-accept=socket,host=localhost,port=2002;urp;"
$ python ooextract.py --pdf filename.doc

The result is almost the same of the one of the easy way but this will use OpenOffice for the conversion, so it will do it better. You also may like to write a little shell script to automate the conversion of a bunch of files, so there it is a very simple version:

#!/bin/bash

openoffice -invisible "-accept=socket,host=localhost,port=2002;urp;"
for i in *.doc; do
	python ooextract.py --pdf $i
done

Remember to kill OpenOffice when it ends :o) OpenOffice has now batteries included.

Google Chrome and proxies

November 11th, 2008 § 0 comments § permalink

I’m paranoid. I don’t want my family be able to use Internet Explorer, at least not from my computer! To be sure they don’t I did one simple thing: set up internet explorer proxy to 1.1.1.1:1 :^D

Unfortunately Chrome won’t work, just like IE, this is because of Chrome and IE share the same proxy configuration. That’s not so good for me, because I like to surf with firefox AND chrome (most of times to see differences between them). My question was simple: how can I set up a different proxy (none in this case) for Chrome and IE?

Chrome accept proxy settings as command line parameters, so you can change the shortcut from something like "C:\Documents and Settings\usr\Local Data\Application Data\Google\Chrome\Application\chrome.exe"

to: "C:\Documents and Settings\usr\Local Data\Application Data\Google\Chrome\Application\chrome.exe" -proxy-server=

that means “do not use proxy at all” :^D

GCJ – Practice Contest – Old Magician

October 13th, 2008 § 0 comments § permalink

Wow, an other practice contest came out :^D here’s the first problem, very simple solution :^)

from __future__ import with_statement
 
def solve(w,b):
    return "BLACK" if b%2==1 else "WHITE"
 
def eatFile(fin, fout):
    with file(fin,'r') as f1:
        N = int(f1.readline().replace("\n",''))
        with file(fout,'w') as f2:
            for case in xrange(1,N+1):
                w,b = [int(i) for i in f1.readline().replace("\n",'').split(' ')]
                s = solve(w,b)
                #print "Case #%d: %s" % (case, s)
                f2.write("Case #%d: %s\n" % (case, s))
 
eatFile('A-large-practice.in','A-large-practice.out.txt')

GCJ 2008 – Round 1A: Minimum Scalar Product

August 5th, 2008 § 3 comments § permalink

This one was really really really too simple. The easiest problem of round 1. Unfortunately I didn’t took part at 1A, I did 1B and 1C and I did not pass :^( Anyway I’ll keep doing those problem just to have some fun :^)

def solve(a,b):
    a.sort()
    b.sort(reverse = True)
    return sum([i[0]*i[1] for i in zip(a,b)])

You have to pass a and b as lists, or if you prefer vectors… I think there is no more to comment here… I did it just thinking about it and I find the solution out in about 3 minutes, but I discovered that mathematicians already thought about that and gave it a name :^D

You can search on wikipedia for Rearrangement inequality. I mean, 4 lines of fully readable python… it was quite simple, wasn’t it?

return YAWN;

Symfony without ssh and pear

August 4th, 2008 § 0 comments § permalink

I’m currently working on a new project and I wanted to study something new to do it. It will be a web application, written in php. No matter what it will be useful for but I’ve found symfony, a great php framework. I read The Book in these 3 days, I’ll start tomorrow developing my app, I should be able to finish the model part in few days, I hope symfony helps :^)

Use it in local is quite simple, there are lot of tutorials and how-to on the symfony site. The painful part is to use it on a common hosting service.

In fact a very powerful tool of symfony is the command line interface, and you usually don’t have ssh support on your hosting service. There is a great tutorial here to make things work, and it’s quite simple, really! I did it today, now it’s working both on the server and on my pc. Look at it ;^)

return ‘bye’;

GCJ 2008: Fly Swatter

July 19th, 2008 § 8 comments § permalink

Consider this:
GCJ 2008 - Problem C - raquet

Believe it or not, but that’s a racquet! In fact the third problem of the Qualification Round of Google Code Jam 2008 is about balls and racquets (or rackets if you prefer)… well… actually it’s about probability :^D

You can find the whole text of the problem here, I’m reporting just the interesting part (IMHO):

Problem

What are your chances of hitting a fly with a tennis racquet?

To start with, ignore the racquet’s handle. Assume the racquet is a perfect ring, of outer radius R and thickness t (so the inner radius of the ring is R-t).

The ring is covered with horizontal and vertical strings. Each string is a cylinder of radius r. Each string is a chord of the ring (a straight line connecting two points of the circle). There is a gap of length g between neighbouring strings. The strings are symmetric with respect to the center of the racquet i.e. there is a pair of strings whose centers meet at the center of the ring.

The fly is a sphere of radius f. Assume that the racquet is moving in a straight line perpendicular to the plane of the ring. Assume also that the fly’s center is inside the outer radius of the racquet and is equally likely to be anywhere within that radius. Any overlap between the fly and the racquet (the ring or a string) counts as a hit.

The input file consists in N lines containing values of f, R, t, r and g separated by spaces.

I’m now trying to explain my approach. The probability we’re looking for it’s given by total_hitting_area/total_racquet_area, or 1 – total_not_hitting_area/total_racquet_area

total racquet area it’s simple:
GCJ 2008 - Problem C - big
I’ll call it big, and it’s given by pi*R^2

The problem is the other area. I’ll try to calculate the not hitting area. Calculus gives a solution to get area of any kind of shape: integrals, but I don’t think it’s necessary to use them in this case. First for all we need to simplify. I don’t care about the ball, I’ll care about the center of the ball. The center of the ball has to be a distance f (the ball radius) from the border of the racquet in order to donot hit it…

This is so the interesting area:
GCJ 2008 - Problem C - small
I’ll call it small, and has a radius of R-t-f so the area is given by pi*(R-t-f)^2

The probability of the center of the ball to be in the small area is given by small/big, and I call it q1

Now consider this:
GCJ 2008 - Problem C - square

The two red areas are identical, they are both composed by the same amount of green and white area, I can consider any of them as a square, its area is given by (g+2*r)^2

Note that small is composed by squares. Some of them are cut off, but that’s should not be a problem… it should

Now consider an hole:
GCJ 2008 - Problem C - gap
I’ll call the red area gap, it’s a little square with sides long g-2*f as I subtract 2 times the ball radius. In fact the center of the ball has to be at least at a distance of f from the sides of the white area in order to do not hit a string. So a gap is (g-2*f)^2

The probability of the center of the ball to pass through a gap contained in a square is given by gap/square and I’ll call it q2

In order to do not hit the racquet to events have to happen at the same time:

    1. The center of the ball has to be in the small area
    2. The center of the ball has to be in a gap area

These to events have probability q1 and q2 respectively.
The probability they are both verified is q = q1 * q2, and that’s the not successful case.

The successful case is p = 1 – q… the probability we were looking for…

It should work… it should… But it doesn’t.

I tried an other approach so. I calculate how many squares there are in the small area. That number is nos = small/square (where nos stands for number of squares)

We know there are as many gap as squares, that’s just because there is one gap inside each square.

In this way I can calculate the total gap area, that is just nos*gap and I call it totalgap

Do you remember the 1 – total_not_hitting_area/total_racquet_area… well the total_not_hitting_area is just our totalgap!

So we have q = totalgap/big… and p = 1 – q

I’ve found the same p as the one above… and it’s wrong.

I went so close the solution but there is something wrong, just to show you, these are the sample cases:
0.25 1.0 0.1 0.01 0.5
0.25 1.0 0.1 0.01 0.9
0.00001 10000 0.00001 0.00001 1000
0.4 10000 0.00001 0.00001 700
1 100 1 1 10

The Google’s results for these cases are:
Case #1: 1.000000
Case #2: 0.910015
Case #3: 0.000000
Case #4: 0.002371
Case #5: 0.573972

But mines are:
Case #1: 1.000000
Case #2: 0.920132
Case #3: 0.000000
Case #4: 0.002364
Case #5: 0.573156

They are so close.

You can download my solution file by clicking here.
That’s my implementation in python:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
def solve(f,R,t,r,g):
    if 2*f >= g: return 1.0
    t += f
    g -= 2*f
    r += f
 
 
    big = pi*R**2
    small = pi*(R-t)**2
    square = (2*r+g)**2
    gap = g**2
 
    # Way 1
 
    q1 = 1.0 * small/big
    q2 = 1.0 * gap/square
    q = q1*q2
 
 
    # Way 2
    '''
    nos = 1.0*small/square
    totalgap = gap*nos
    q1 = 1.0 * small/big
    q2 = 1.0 * totalgap/small
 
    q = q1*q2
    '''
 
 
    if 0 <= q:
        if q <= 1:
            p = 1-q
        else:
            p = 0.0
    else:
        p = 1.0
    #print 1-q
    return p
 
 
def sfs(s):
    'solve from string'
    s = s.replace('\n', '').split(' ')
    f,R,t,r,g = [float(i) for i in s]
    return solve(f,R,t,r,g)

I did one in C too, I was thinking python did something wrong with math…

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
#include <stdio.h>
#define pi 3.1415926535897931
 
double solve(double f,double R,double t,double r,double g)
{
    double big,small,square,gap,q1,q2,q,p;
 
    if (2*f >=g )
    {
         return 1;
    }
 
    t = t + f;
    g = g - 2*f;
    r = r +f;
 
    big = pi*R*R;
    small = pi*(R-t)*(R-t);
    square = (2*r+g)*(2*r+g);
    gap = g*g;
 
    q1 = small/big;
    q2 = gap/square;
    q = q1*q2;
 
    if (0<=q)
    {
        if (q<=1)
        {
            p = 1-q;
        } else p = 0;
    } else p = 1;
 
    return p;
}
 
int main()
{
    printf("Case #1: %lf\n",solve(0.25, 1.0, 0.1, 0.01, 0.5));
    printf("Case #2: %lf\n",solve(0.25, 1.0, 0.1, 0.01, 0.9));
    printf("Case #3: %lf\n",solve(0.00001, 10000, 0.00001, 0.00001, 1000));
    printf("Case #4: %lf\n",solve(0.4, 10000, 0.00001, 0.00001, 700));
    printf("Case #5: %lf\n",solve(1, 100, 1, 1, 10));
}

There’s no need to say that the result was the very same.

I don’t know what’s wrong with my approach, all the others’ code I read used something involving integrals, sometime already solved and they just did the “definite” part.

Did you read someone’s code that is not using integrals but pure geometry and probability approach?
I think it’s possible, I think there is something very bastard I’m missing… don’t know what… do you?

return “grr”

GCJ 2008: Train Timetable

July 19th, 2008 § 0 comments § permalink

This one was the title of the second problem of Google Code Jam Qualification Round 2008.

This time my code is a bit messy, but I won’t adjust it for you :^D It’s the disadvantage of coding fast, at least for me. I often wrote something not necessary, someone call it gold planting, it’s just what I did in the maze problem with the draw method…

You can download my solution for the Train Timetable problem I wrote.

Here is the commend code.

1
2
from __future__ import with_statement
from copy import copy

Don’t ask me why I imported copy, I think I did it while testing… maybe. I don’t remember it’s not useful
The first 3 are classes but I did not spent so much time projecting their structure. Tempo just means time in italian, I didn’t want to override the python name (I know that’s the name of a lib, it’s only an excuse, I’m italian, let me name things in italian sometime :^P)

tempo is not so useful in that form… anyway converts the format ‘hh:mm’ to an int representing minutes

4
5
6
7
8
class tempo(object):
    def __init__(self, s):
        a = s.split(':')
        x = [int(i) for i in a]
        self.m = x[0]*60 + x[1]

Trip represents a trip, with starting and ending time.

11
12
13
14
15
16
17
18
19
20
21
22
23
24
class Trip(object):
    def __init__(self,start,end):
        self.start = tempo(start)
        self.end = tempo(end)
        self.duration = self.end.m - self.start.m
        self.busy = False
    def __cmp__(self,x):
        return cmp(self.start.m,x.start.m)
    def use(self):
        self.busy = True
    def isfree(self):
        return self.busy == False
 
    is_free = property(isfree)

There is no need to put the duration attribute, but it did, thinking on it now it’s pretty useless… The interesting part is the flag busy, or is_free (just a joke with python property function…). A Trip is free if you did not use it yet. I added __cmp__ method because I’ll need to “sort” them, if trip X starts earlier than trip Y well…

X<y == True

It will be useful to understand what’s the first Trip to use…

A Train is identified by the time of the last thing it did (self.Time) and it has done something or not (self.init)

28
29
30
31
32
33
34
35
36
37
38
39
40
class Train(object):
    def __init__(self,Time):
        self.Time = tempo(Time)
        self.init = False
    def TripTrough(self, trip, turnaround):
        'Returns True if the train can trip (and it does)'
        if (trip.is_free and self.Time.m+turnaround <= trip.start.m) or self.init ==False:
            self.init = True
            self.Time = trip.end
            trip.use()
            return True
        else:
            return False

The TripTrough method is very important (yes I know, it’s Through not Trough… my english sucks, sorry for that :^D )… Anyway… How does it work? It’s simple. It wants a trip and the turnaround time. the flag init it’s useful because if the trip starts at 0:03 and you have a turnaround of 5 you have to don’t care about the turnaround, you have no need to turn around… my check can’t work without an init flag… I don’t know if it’s clear..

Now something magical happens, I’m doing it bad I think, I’m playing with the “all name are references” of python… I changed the trip flag as the train used it, I could do it out of that method but maybe I was tired, bored, drunk, whatever… and I did it in this way :^D. I updated the time of the train too, obviously.

Table has the algorithm! Let me explain…

45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
class Table(object):
    def __init__(self,A,B,T): # trip lists... and turnaround
        self.A = A
        self.B = B
        self.T = T # turnaraund
    def solver(self, starting,ending,train):
        starting.sort()
        s = starting
        for i in s:
            if train.TripTrough(i,self.T):
                return self.solver(ending,starting,train)
        # else...
        return train
 
 
    def solve(self):
        tfa = []
        tfb = []
 
        check = True
        while check:
            freeA = [i for i in self.A if i.is_free]
            freeB = [i for i in self.B if i.is_free]
            #print freeA, freeB
            if len(freeB) > 0 and len(freeA)>0:
                mins = min(freeA),min(freeB)
            elif len(freeB) == len(freeA) == 0:
                break
            if len(freeA)>0 and (len(freeB) == 0 or mins[0]<mins[1]): #start from A
                tfa.append(self.solver(freeA,freeB,Train('00:00')))
            else:
                tfb.append(self.solver(freeB,freeA,Train('00:00')))
        return len(tfa),len(tfb)
  • A is the list of trips starting from station A going to B
  • B is the list of trips starting from station B going to A
  • T is the turnaround time of the case
  • solver it’s a recursive method.
    It takes 3 parameters:

    1. starting it’s the list of the trips starting from a station
    2. ending it’s the list of the trips starting from the other station
    3. train is the train that’s traveling forward and backward

    Well… I sort the trip list, in this way come first the one that will start first, then I begin to look for the first one that the train can travel through. If it can I can look (recursively part) for a return trip, just inverting starting with ending calling once again the solver method.

  • solve is the method you need to call to get the final result.
    • tfa is the list of trains starting from the A station
    • tfb is the list of trains starting from the B one
    • The part
      65
      66
      
              check = True
              while check:

      it’s just useless… I could use while True… I won’t change check

    • freeA is the list of the trip starting in A that are not busy
    • freeB it’s the same as the above one but in B, obviously
    • if there is something in freeA and freeB I’d like to know what’s the trip starting earlier in freeA and the one in freeB
    • else if there is nothing in freeA and nothing in freeB there is no need to continue
    • if there is something in freeA and there is nothing in freeB or just the earliest trip in freeA starts before the earliest one in freeB I can set up a new train and make it travel starting from the earliest trip in freeA until it stops…
    • else I can set up a new train and make it travel starting from the earliest trip in freeB until it stops…
    • I added these trains to tfa and tfb
    • I return the number of items of tfa and tfb, as the problem asks..

You can download the solution here with the file processing too.

The algorithm it’s correct but the implementation sucks, I think I can code a better one with less line of code, I’ve just to spend some time on it… It was a time challenge and I was not thinking so much of optimization… I don’t know If I’ll post a revisited one… For now it’s all folks!

return ‘Bye’