Doc Gecko

Creating Facebook Account and Page access tokens for the fb_graph gem

2011-12-22T00:00:00-08:00

Creating Facebook Account and Page access tokens for the fb_graph gem

22 Dec 2011 – Lisboa | rails fb_graph gem facebook

I have a jobs board app built in Ruby on Rails (http://infrajobshq.com). As jobs are posted to the site by job advertisers, the app posts shortened information about each jobs to both Twitter and Facebook pages.

Recently, the app started failing when jobs were posted and I wasn’t sure why. After investigation, I found that it was because I had changed by Facebook password. I didn’t realise that the password is also associated with the ACCESS_TOKEN of Facebook Graph API.

So, again I needed to go through the process to generate the ACCESS_TOKEN for my app. As such, I thought I would post the process as it requires a little bit of work and is not necessarily an easy process to remember: I had to go through a number of pages on the Facebook Developers section in the Facebook Graph API (https://developers.facebook.com/docs/reference/api/).

The process is a 2 step process. Firstly you have to obtain an ACCESS_TOKEN that gains you access to your Facebook account. Once you have that level of access, you then need to find the access token of your Facebook Page you would like to gain access to. So here goes:

1. Step 1: Obtain Facebook Account Access Token

In order to gain your Facebook account ACCESS TOKEN, you can use the following url (in your chosen browser):

https://www.facebook.com/dialog/oauth?
     client_id=YOUR_APP_ID&redirect_uri=YOUR_URL&scope=manage_pages&
     response_type=token

YOUR_APP_ID and YOUR_URL can be found on your personal Facebook Developers page, where you apps exist, i.e.:

https://developers.facebook.com/apps

As you can see from this page, under the Summary section, you can find your:

App ID/API Key
Site URL

So, substituting YOUR_APP_ID with your “App ID/API Key” and YOUR_URL with your “Site URL”, you will have a url that looks something like this:

https://www.facebook.com/dialog/oauth?
     client_id=542148432111423&redirect_uri=http://infrajobshq.com/&scope=manage_pages&
     response_type=token

As you enter this url, you will be redirected to your site’s url, in my case to http://infrajobshq.com. And now, as you will see from the url, it contains additional information after sites main address, e.g. it looks something like this:

http://infrajobshq.com/#access_token=BBBDGIb5OYp4BAJK0qYw8qw3fY4uPps23xfMpvobHvVCN4OsI50LGIshL8fUZA
		 xQsgnUt45b6kAgtFNm3X4wezyvK7VWEZD&expires_in=0

As you can see, the url contains the ACCESS TOKEN for your Facebook Account. You will need this for the next step.

2. Step 2: Obtain Facebook Page Access Token

Now, to find the ACCESS TOKEN for your Facebook Page, use the following url:

https://graph.facebook.com/me/accounts?access_token=ACCESS_TOKEN

Simply replace the ACCESS_TOKEN with your Facebook Account ACCESS_TOKEN (and no need to include the bit on the end of the url that is ‘&expires_in=0’). In my case, the url looks like this:

https://graph.facebook.com/me/accounts?access_token=BBBDGIb5OYp4BAJK0qYw8qw3fY4uPps23xfMpvobHvVCN4O
		 sI50LGIshL8fUZAxQsgnUt45b6kAgtFNm3X4wezyvK7VWEZD

Once you have entered this, you will be given a page that contains your Facebook Page ACCESS_TOKEN, e.g.:

{
   "data": [
      {
         "name": "InfraJobs",
         "access_token": "BBBDGIb8OUp4BDNWFlERsVcEaeS1lylDAqpN3FtDET4L2fdeS6TPoNvYZAZBbzeDXC6TKTB8w
  LnX3wPZA0HU8n6pvZCmRndGwOcQmvU1T4BAZDZD",
         "category": "Consulting/business services",
         "id": "130231394756745"
      },
      {
         "name": "InfraJobs",
         "access_token": "AAADGIb8OYp4BAMJBpufj6LhsDq2W8JhCpQW6P1zjZCK32uwuK0njrRItJybbfFPXKEOsZCCf
  4Sfjxkmk4dcJz43OQN8ttOKoPZClOnJZCwZDZD",
         "category": "Application",
         "id": "327845241634422"
      }
   ],
   "paging": {
      "next": "https://graph.facebook.com/me/accounts?access_token=BBBDGIb5OYp4BAJK0qYw8qw3fY4uPps2
  3xfMpvobHvVCN4OsI50LGIshL8fUZAxQsgnUt45b6kAgtFNm3X4wezyvK7VWEZD&limit=5000&offset=5000&_
  _after_id=327845241634422"
   }
}

As you see, there are two sets of data related to my site. Once relates to my Facebook Page and the other relates to the Facebook App of the same name. The ACCESS_TOKEN you require is the one related to the Facebook Page: in my case its the the first once in the list.

Wow, so many ACCESS_TOKENs hey!

Anyway, I am using these codes in my App to post to my Facebook Page, so my Ruby on Rails code look like this:

def facebook(message)
  fb_message = message(message) + ": " + job.short_url
  me = FbGraph::User.me("BBBDGIb5OYp4BAJK0qYw8qw3fY4uPps23xfMpvobHvVCN4OsI50LGIshL8fUZAxQsgnUt45b6k
	AgtFNm3X4wezyvK7VWEZD")
  me.feed!(
    :access_token => "BBBDGIb8OUp4BDNWFlERsVcEaeS1lylDAqpN3FtDET4L2fdeS6TPoNvYZAZBbzeDXC6TKTB8wLnX3
		wPZA0HU8n6pvZCmRndGwOcQmvU1T4BAZDZD",
    :message => fb_message
  )
end

As you can see, the Facebook Account ACCESS_TOKEN is used in the 3rd line of code, and the Facebook Page ACCESS_TOKEN is used in the 5th line of code. The first one gains me access to my account and the second that allows access to the the Page’s feed.

Damn, finally got there. I realise now that I was a little confused the first time I did this, simply as there were a few ACCESS_TOKENs knocking about, and that I wasn’t initially clear from the Facebook Developer Graph API Authentication which urls I should have been using. But now after going through the process, its quite straight-forward. Even so, glad its all written down here now :-).

time.strftime(string) reminder

2011-06-23T00:00:00-07:00

time.strftime(string) reminder

23 June 2011 – Lisboa | rails strftime reminder

This is a major reminder, as I constantly have to search for this one whenever it comes to formatting dates and times in Ruby / Rails:

%a - The abbreviated weekday name (``Sun'')
%A - The  full  weekday  name (``Sunday'')
%b - The abbreviated month name (``Jan'')
%B - The  full  month  name (``January'')
%c - The preferred local date and time representation
%C - Century (20 in 2009)
%d - Day of the month (01..31)
%D - Date (%m/%d/%y)
%e - Day of the month, blank-padded ( 1..31)
%F - Equivalent to %Y-%m-%d (the ISO 8601 date format)
%h - Equivalent to %b
%H - Hour of the day, 24-hour clock (00..23)
%I - Hour of the day, 12-hour clock (01..12)
%j - Day of the year (001..366)
%k - hour, 24-hour clock, blank-padded ( 0..23)
%l - hour, 12-hour clock, blank-padded ( 0..12)
%L - Millisecond of the second (000..999)
%m - Month of the year (01..12)
%M - Minute of the hour (00..59)
%n - Newline (\n)
%N - Fractional seconds digits, default is 9 digits (nanosecond)
        %3N  millisecond (3 digits)
        %6N  microsecond (6 digits)
        %9N  nanosecond (9 digits)
%p - Meridian indicator (``AM''  or  ``PM'')
%P - Meridian indicator (``am''  or  ``pm'')
%r - time, 12-hour (same as %I:%M:%S %p)
%R - time, 24-hour (%H:%M)
%s - Number of seconds since 1970-01-01 00:00:00 UTC.
%S - Second of the minute (00..60)
%t - Tab character (\t)
%T - time, 24-hour (%H:%M:%S)
%u - Day of the week as a decimal, Monday being 1. (1..7)
%U - Week  number  of the current year,
        starting with the first Sunday as the first
        day of the first week (00..53)
%v - VMS date (%e-%b-%Y)
%V - Week number of year according to ISO 8601 (01..53)
%W - Week  number  of the current year,
        starting with the first Monday as the first
        day of the first week (00..53)
%w - Day of the week (Sunday is 0, 0..6)
%x - Preferred representation for the date alone, no time
%X - Preferred representation for the time alone, no date
%y - Year without a century (00..99)
%Y - Year with century
%z - Time zone as  hour offset from UTC (e.g. +0900)
%Z - Time zone name
%% - Literal ``%'' character

t = Time.now                        #=> 2007-11-19 08:37:48 -0600
t.strftime("Printed on %m/%d/%Y")   #=> "Printed on 11/19/2007"
t.strftime("at %I:%M%p")            #=> "at 08:37AM"

Kill rails server

2011-05-27T00:00:00-07:00

Kill rails server

27 May 2011 – Lisboa | rails grep console

This is just a brief reminder to myself of how to get rid of an already running rails server (for when it happens).

All you need to do in the console, within your application subdirectory is type:

ps aux | grep rails

After which you should see some output like this:

Daren    89314   0.2  2.0  2551940  82732 s003  R+    2:38PM   2:38.93 /Users/Daren/.rvm/rubies/ruby-1.9.2-p180/bin/ruby script/rails s
Daren     2881   0.1  0.1  2524112   3296 s005  R+   Wed06PM   2:11.62 /Users/Daren/.rvm/rubies/ruby-1.9.2-p180/bin/ruby script/rails c

The second line is my running server that I need to shutdown, which can be done as follows:

kill -9 <pid>

In my case, the pid is 89314, so I use the following:

kill -9 2881

You can check if the server is down, by again typing in:

ps aux | grep rails

In my case, I would see something like this:

Daren     2881   0.1  0.1  2524112   3296 s005  R+   Wed06PM   2:11.62 /Users/Daren/.rvm/rubies/ruby-1.9.2-p180/bin/ruby script/rails c

i.e the rails server process has been killed!

Upgrading and downgrading your rubygems

2011-04-10T00:00:00-07:00

Upgrading and downgrading your rubygems

10 Apr 2011 – Lisboa | rails rubygems sunspot

I just upgraded to rubygems version 1.8.1, which generates a large number of depreciation errors. This is not a major problem as all still continues to work fine, expect for a couple of my gems: the sunspot (1.2.1) and sunspot_rails (1.2.1) gem. Currently, under gem 1.8.1, I was unable to start or reindex solr.

As such, for now I decided to revert back to version 1.7.2 of rubygems. In order to downgrade your rubygems version, you have to do the following:

gem uninstall rubygems-update -v 1.8.1
gem install rubygems-update -v 1.7.2
update_rubygems _1.7.2_

And, going the other way, if you wish to update to the latest version, simply run:

gem update --system

Excel SUMIF with criteria in multiple columns

2010-12-01T00:00:00-08:00

Excel SUMIF with criteria in multiple columns

1 Dec 2010 – Lisboa | excel sumif multiple criteria

The SUMIF function is designed to SUM one column if another column contains the some expected criteria. For those who have already noticed, the SUMIF function does not work when your criteria exists in multiple columns. This is an issue I just had to face.

Multiple Column Criteria

When criteria exists in more than 1 column, you have to take a completely different approach, e.g. when you have a situations such as given in the table below:

Here I have 3 columns:

Column A contains a Length (in meters).
Column B contains a Code (i.e. 2, 4a & 4b)
Column C contains a Score (1 to 4)

I need to check that both data in Columns B and C match the criteria of specific Code and Score, and for those that match both criteria, I need to SUM the lenths in Column A.

To do this you need to use an Array Formula, also known as a CSE Formula or even a Super Formula. This is a formula that requires you to press Ctrl+Shift+Enter after typing it, instead of just Enter.

Example

If you wanted to sum together all Lengths where the Code is 4a and the Score is 2, here is the formula you can to use:

=SUM((B2:B11="4a")*(C2:C11="2")*A2:A11)

This will find that rows 4 and 11 contain both criteria at the same time, and will thus return the sum of lengths in those columns, i.e. A4 and A11, which is 141.09m + 731.38m = 872.47m.

The formula looks a little strange and probably not what you’d expect; however, it works very well, as long as you remember to use Ctrl+Shift+Enter instead of Enter when inputting the formula.

Excel cell validation - counting of non-blank cells

2010-08-01T00:00:00-07:00

Excel cell validation – counting of non-blank cells

1 August 2010 – Lisboa | excel validation countblank

When building Excel-based models, more-and-more it is becoming clear that it is necessary to include such functionality as Validation and, to some degree, to the degree expected from web-based software.

With a model I am building at the moment, it has been necessary to reduce the chance of a user not including certain data that would consequently cause the model to crash when run. As such, to reduce the likelihood of not entering all of the required data, I am building in checks that will then provide an error list if certain data is not entered.

Error Checking

The error checking includes a count of remaining errors in an obvious place on the input excel sheet, i.e. in a place that can be easily checked by the user.

So my problem was that I need to firstly check each input to see if data has been entered or not. So firstly I have a single column of cells (e.g. column A3:A()) that must be filled with data by the user. In the column next to this (e.g. column B3:B8) is the data check. This column contains an initial data check, i.e. Cell B3 contains the following (which can be copied down the other cells in column B):

=IF(AND($A3<>""),"Please enter data", "")

So if the user fills in cell A3, B3 remains blank, i.e. it does not report an error. However, if A3 does not contain data, the error appears in Cell B3 saying “Please enter data”

Error Count

So my actual problem was now to count the number of remaining errors, i.e. how to count the number of cells in column B that report the error. Sounds easy to do right…well, thats what I thought. I thought I have to do is count the number of NON BLANK cells in column B. Well, for example using any of the following do not work:

=COUNTIF(B3:B8,<>"") or =COUNTIF(B3:B8,"<>""") or =COUNTIF(B3:B8,"<>")

In the end I used an approach offered by Prasanth Chandra. What you have to do is essentially count all of the cells in the range and remove from that the number of blank cells, as follows:

=(ROWS(B3:B8)*COLUMNS(B3:B8))-COUNTBLANK(B3:B8)

Looks complex, but works right away!

Welcome to Doc Gecko's Blog

2010-07-11T00:00:00-07:00

Welcome to Doc Gecko’s Blog

11 Jul 2010 – Lisboa | asset risk performance modelling

I began modelling the performance of assets as part of my PhD, when I developed a methodology for predicting the life of an ultra-reliable electro-actuator for deepwater oil & gas production, for Rotork Actuation Ltd. The results identified a number of design-risks that were subsequently modified in advance of actually building the actuator.

However, since 1999 I have been developing Asset and Performance-based Risk Models for commercial use, with great success and varied / continual use on both Private Finance Initiative (PFI) and Public Private Partnership (PPP) projects, including:

Dutch (Southern) High Speed Rail link
Tube Lines Ltd
Dutch Gelre line Tender
M25 DBFO (Design Build Finance Operate)
Crossrail

There varied uses have included:

Prediction of the Performance of Assets over a 30-year period (Tube Lines Ltd)
Expert Maintenance Optimisation of a 7-year infrastructure rail concession (Spitzke Spoorbouw, S.A.)
Whole Life Costing of design options, over a 60-year period (Crossrail)
Lane Availability Prediction of the UK M25 Motorway

Unlike the electro-actuator model, which was based on a particular failure mechanism (and thus at component level), the later models have been based on modelling entire infrastructure systems, such as Rail and Highways networks.

Furthermore, instead of focusing on on asset failure alone, they have linked failure, degradation and operations to the cost of interventions and contractual requirements, as found in many PPP and PFI projects, including costs of:

Major Capital Works (Modernisations and Upgrades)
New Infrastructure
Minor Upgrades
Planned and Reactive Maintenance
Performance-Penalties & Bonuses
Operational Modifications (e.g. change to train routes and time-tabling)

The most recent models have been built using a combination of Microsoft Excel, Visual Basic for Application (VBA), Visual Basic (VB6+) and Palisade’s @Risk.

But now I am in the process of transferring the modelling approach to Ruby on Rails.

As such, this blog will discuss both the Excel and Rails approaches to modelling, as well as a few other bits-n-bobs!