Piping
PIPING in CLAN refers to the facility for directing the output from one analytic tool (for example COMBO or GEM) to another tool for a second analysis. The output from the combined analyses is then collated into a single output file. For example, should you have annotated a batch of transcripts with a particular GEM collection, and would like to conduct further searches within this collection and this collection only, the GEM command can be exported to a second tool, which will conduct an analysis on the particular GEM sequences, before generating an output file.
We will look at this, while referring to the Ex11_PIPING .cha file in the workshop folder.
This is how:
- In the workshop folder there is a PIPING exercise file. Open it.
- You see 2 GEM collections: ‘laughter_sequence’ and ‘other_repair’1. Two of the ‘repair’ GEMs are embedded within ‘laughter’ sequences, and one ‘repair’ GEM is not.
- We are only interested in the ‘repair’ GEMs that feature in the ‘‘laughter_sequence’ GEM sequences, and are going to generate a file that includes these, and excludes the other one.
- Open commands, and type
- gem +slaughter_sequence
- +d (this directs the programme to export the collection as a CHAT file)
- File in> choose the file you are subjecting to analysis (Ex13_PIPING)
- | (the piping symbol)
- gem +sother_repair2
- in full, it should look like this
gem +slaughter_sequence +d @ | gem +sother_repair
This basically denotes that an initial analysis will be conducted and will output all the
‘laughter_sequence’ GEMs, it will furthermore output these as a .cha file (+d), and this will then will be piped (|) to a second analysis, where only the GEM sequences marked ‘other_repair’ will be output.
The eventual file will then only include the 2 ‘other_repair’ sequences that were embedded within the ‘laughter_sequence’ section, and will exclude the ‘other_repair’ GEM that was not associated with a ‘laughter_sequence’ sequence.
See below for output file. You see that it includes the 2 ‘other_repair’ sequences embedded in the initial ‘laughter_sequence’ sequence, and has not included the third ‘other_repair’ GEM.
Remember to save this output file as a .cha file if you want to play the data from this file.
EXERCISE
Conduct an analysis on the laughter_sequence GEM collection, specifically looking at how much each participant speaks (hint: the duration of each person’s time at talk within these sequences).
1 It should be noted that the workshop files are for training purposes only, and do not include annotations based on proper analysis
2 Note that the programme only searches for the initial standalone word, unless directed not to (see info sheet on COMBO for directions on how to combine words in a search)
> gem +slaughter_sequence +d @ | gem +sother_repair
gem +slaughter_sequence +d @
Fri Jan 9 15:29:08 2015
gem (20-Sep-2013) is conducting analyses on:
ALL speaker tiers
and ONLY header tiers matching: @BG:; @EG:;
****************************************
From file </Users/spencer/Workshop CLAN/Nottingham/02. Linking Software Workshop
Exercises/Ex11-Piping.cha>
From file </Users/spencer/Workshop CLAN/Nottingham/02. Linking Software Workshop
Exercises/JMSH_BLT_BV_20130513_YT_IN.cha>
gem +sother_repair
Fri Jan 9 15:29:08 2015
gem (20-Sep-2013) is conducting analyses on:
ALL speaker tiers
and ONLY header tiers matching: @BG:; @EG:;
****************************************
From pipe input
*** File “pipeout”: line 12.
@BG: other_repair
*DIM: but five of them have been translated into english
(0.5)
*ROS: or six
*FRA: well three
*ROS: three ⁇co-⁇ erph- ⌈i’ve got it i’ve got quite a long list here⌉
@EG: other_repair
@EG: laughter_sequence
*** File “pipeout”: line 26.
@BG: other_repair
*FRA: ⌊i do love⌋ sailing boats
*ROS: you d- you love sailing boa⌈ts⌉
*FRA: ⌊no ⁇b-⌋ no⁇ just kidding
@EG: other_repair
@EG: laughter_sequence
*** File “pipeout”: line 36.
@BG: other_repair
*ROS: i- i think people think that germans eat a lot of mus⌈sels⌉
*BIR: ⌊no⌋
not at all
*ROS: no↗
@EG: other_repair
@EG: laughter_sequence