count questions and (max) points per category (Feature #91)
Description
It would be help full to be able to pick up the total number of questions in an exam as well as the maximum points per category.
Our exam practice often has two question categories, general, mc questions, to be shuffled each question with typically the same max number of points per question and open questions, not shuffled and typically a different number of points.
It would be nice to be able to count the questions and the maximum number of points per category to be able to use this information to be put on the cover page of the exam.
Our regulations stipulates that we put this information on our cover page.
Probably these counters are already implemented but not documented for public use.
History
Updated by Alexis Bienvenüe over 11 years ago
AMC processes the scoring strategy and all scoring computations within the perl programs, not with LaTeX. LaTeX is only used to pass to the perl program the scoring strategies and the wrong/right answers. So I'm afraid the only solution is to hard-code the maximum number of points in the LaTeX source, and maintain the correspondence between this value and the scoring strategy...
Updated by Pieter Van den Hombergh over 11 years ago
Then probably some perling on my side can also help.
Since I can either infer from the questions how many points they get or add an annotation per question file. (As some kind of structured comment as in %%maxpoints=3)
Since we put the reference to questions in separate file per category, like
\element{general}{\inputQ{02threadsafety/oProperGuarding.tex}}
\element{general}{\inputQ{02threadsafety/oThreadSafeCode1.tex}}
summing the points should be easy enough.
Is it possible to add a pre-update hook in the process cycle?
This pre-update hook would be a command/script, specified either in the global settings or per project.
It should be invoke just before the latex processor (like pdflatex). This is what I would do if I would use a make file to compile the text.
This could then easily run such a script as mentioned above before the latex run.
Such script would produce a file with a well known name, such as counts.tex, which then again could be inputted (or \InputIfFileExists{}) in the preamble, et voilà, we have what we need.
Find simple script attached. It is just a hack, fitting to our practice.
- File countpoints.pl added
Updated by Alexis Bienvenüe over 11 years ago
The problem is difficult to solve for all exams, because the scoring strategies are allowed to be quite complex, and the questions given to different students may vary, as well as the maximum scores for a given question...
I can see two possible solutions for your problem, although I must admit none of these are very good.
1. Using LaTeX
Define yourself the counter to be initialized at the beginning of a copy, and add the max scores for each question in turn. You may define some LaTeX macros to get a simpler LaTeX code...
2. Using perl
If you know how to process the source file from a perl program to get what you need, perhaps you can declare a new "source file format" as a plugin - see a very quick howto at Build a new filter plugin.
- Category deleted (
LaTeX)
Updated by Pieter Van den Hombergh over 11 years ago
Alexis Bienvenüe wrote:
The problem is difficult to solve for all exams, because the scoring strategies are allowed to be quite complex, and the questions given to different students may vary, as well as the maximum scores for a given question...
I can see two possible solutions for your problem, although I must admit none of these are very good.1. Using LaTeX
Define yourself the counter to be initialized at the beginning of a copy, and add the max scores for each question in turn. You may define some LaTeX macros to get a simpler LaTeX code...
2. Using perl
If you know how to process the source file from a perl program to get what you need, perhaps you can declare a new "source file format" as a plugin - see a very quick howto at Build a new filter plugin.
Your proposal would lead some kind of generic solution, which might be quite complex. Actually, my requirements are rather more humble. I would like to avoid to have to specify the question count and the
maximum total score, when they should theoretically be computable from the input. If LaTeX does not maintain such numbers during a run, then summation of the score and counting the questions could be done in
an external program, e.g. in perl.
Our way of creating exams uses a fixed set of files, in a directory specific for one exam (very much as AMC does itself)
The files are called source.tex and questions.tex for mc-only exams. The source file defines so meta data on the exam such as date and time.
The questions.tex file only contains lines as above \element{general}{\inputQ{02threadsafety/oThreadSafeCode1.tex}}, which is easily
parsed and processed by the custom perl script that knows of these files and also can derive where the actual question files can be found, in this case for instance, where the subdir 02threadsafety can be found.
I really have no reason to define another file type, I am quite glad with LaTeX.
What would help is somehow run the mentioned perl script as part of the latex processing, so you will not forget. Now I have to run the script by hand every time I add or change a question file. This is not a big deal, but running it from the AMC process would be simple. I trick came up my selves is make a pdflatex wrapper, an executable script that first calls the perl-script and after that the normal pdflatex.
Updated by Alexis Bienvenüe over 11 years ago
I really have no reason to define another file type, I am quite glad with LaTeX.
The file type I proposed to you was a virtual file type, like "LaTeX with a small preprocessing"...
I added a preprocessing option in LaTeX source files from revision r1216: add the line
%%AMC:preprocess_command=countpoints.pl
to your LaTeX source (with no non-comments lines before this one). This will tell AMC to call
countpoints.pl DOC-filtered.tex
from the project directory, after having copied
source.tex to DOC-filtered.tex (so that your command may change the contents of DOC-filtered.tex with no problem).You can write the full path of the command if it is not in the project directory.
Updated by Michaël Cadilhac about 1 month ago
FYI, I did something similar, but in LuaTeX. I have a small Lua script that parses the .amc file at '\end{document}' and produces a list of scores and some more info. This is then put in the .aux file, and when reran, the document has access to all the info.