[Halld-online] [Clas12_software] EVIO dictionaries

[Elliott Wolin]

Hi Mac,

The hierachical scheme is from Maurizio.  The parser generates composite 
names based on the location of the bank or leaf tag in the XML 
hierarchy.  The xpos bank is a child of DC, which is a child of CLAS12.  
Thus the entry in the internal dictionary map uses the name 
"CLAS12.DC.xpos."  It is just a string, i.e. the internal dictionary is 
just a map< string, pair<int,int> >.

I could easily change the composite string scheme, e.g. use "/" instead 
of ".",  or make the separator character user-settable, or anything else 
that makes sense.  I'm open to suggestions.



On 04/16/2012 01:36 PM, Mac Mestayer wrote:
> Hello Elliott;
>
> My question wasn't about your use or mis-use of a particular
> tag name, like "DC"; it was about the dimensionality of the
> tag name.  Is the name just a long string, like "dc-reg2-sec4",
> or does it have a definite structure, like bank=dc, leaf=reg2,
> sub-leaf=sec4 ?
>                 - Mac
>
> "mestayer at jlab.org", (757)-269-7252
>
> On Mon, 16 Apr 2012, Elliott Wolin wrote:
>
>> Hi Mac,
>>
>> Ignore any correspondence between my example and real life, I just 
>> typed whatever tag names came into my head.
>>
>>
>>
>> On 04/16/2012 01:22 PM, Mac Mestayer wrote:
>>> Hello Elliott;
>>>
>>> I liked your examples, but I have a question.
>>> I'm not sure what "DC" and then "xpos" means; we have
>>> 18 drift chambers arranged in 3 'regions' for each of
>>> six 'sectors'.  Don't we need at least one more layer
>>> between 'bank' and 'leaf'?
>>>
>>>             - Mac
>>>
>>> "mestayer at jlab.org", (757)-269-7252
>>>
>>> On Mon, 16 Apr 2012, Elliott Wolin wrote:
>>>
>>>> Hi,
>>>>
>>>> This morning I checked into evio-4.0 a new version that implements a
>>>> dictionary facility that I believe meets all requirements that were
>>>> discussed last week.  To my knowledge this completes the initial round
>>>> of C++ upgrades to EVIO, and now the Java and C++ versions should be
>>>> completely compatible.
>>>>
>>>> A number of other improvements have been discussed, including adding
>>>> append mode and random-access I/O.  We need to discuss with the DAQ
>>>> group which are feasible and set priorities.
>>>>
>>>> I parse three types of XML dictionary tags:  "dictEntry", "bank" and
>>>> "leaf".  In all cases the only attributes I look at are "name", "tag"
>>>> and "num."  All other tags and attributes are ignored, so the XML 
>>>> can be
>>>> used to define whatever else you want, e.g. definitions of composite
>>>> bank structures, content specifications (e.g: contentType="int"), etc.
>>>>
>>>> dictEntry is the simplest, it just makes a entry into the map of names
>>>> vs tag/num pairs.  It can occur anywhere in the XML:
>>>>
>>>> <dictEntry name="fred"  tag="1" num="1"/>
>>>>
>>>> Here the name "fred" is a synonym for the tag/num pair (1,1).
>>>>
>>>> "bank" and "leaf" are for describing hierarchical bank structures, 
>>>> e.g:
>>>>
>>>> <bank name="CLAS12" tag="1" num="0">
>>>> <bank name="DC" tag="20" num="0">
>>>> <leaf name="xpos" tag="20" num="1"/>
>>>> <leaf name="ypos" tag="20" num="2"/>
>>>> <leaf name="zpos" tag="20" num="3"/>
>>>> </bank>
>>>> <bank name="SC" tag="30" num="0">
>>>> <leaf name="xpos" tag="30" num="1"/>
>>>> <leaf name="ypos" tag="30" num="2"/>
>>>> <leaf name="zpos" tag="30" num="3"/>
>>>> </bank>
>>>> </bank>
>>>>
>>>> where "bank" means bank-of-banks, and "leaf" means bank-of-ints (or
>>>> floats or another primitive types).  The parser will generate map
>>>> entries that equate the tag/num pair (1,0) to "CLAS12", (20,0) to
>>>> "CLAS12.DC", (20,1) to "CLAS12.DC.xpos", (30,3) to 
>>>> "CLAS12.SC.zpos", etc
>>>> (actually I think everything is case-insensitive, have to check).
>>>>
>>>> We can discuss other tag names besides "bank" and "leaf," but I think
>>>> these are adequate.  Currently if a leaf entry has children an 
>>>> exception
>>>> is thrown.  I could change this to a warning if desired.
>>>>
>>>> Please give it a try.
>>>>
>>>> Thanks,
>>>>
>>>> -- 
>>>>
>>>>                 Sincerely,
>>>>                     Elliott
>>>>
>>>>
>>>> ================================================================================ 
>>>>
>>>>
>>>>  Those raised in a morally relative or neutral environment will hold
>>>>             no truths to be self-evident.
>>>>
>>>>
>>>> Elliott Wolin
>>>> Staff Physicist, Jefferson Lab
>>>> 12000 Jefferson Ave
>>>> Suite 8 MS 12A1
>>>> Newport News, VA 23606
>>>> 757-269-7365
>>>>
>>>> ================================================================================ 
>>>>
>>>> _______________________________________________
>>>> Clas12_software mailing list
>>>> Clas12_software at jlab.org
>>>> https://mailman.jlab.org/mailman/listinfo/clas12_software
>>>>
>>
>> -- 
>>
>>                 Sincerely,
>>                     Elliott
>>
>>
>> ================================================================================ 
>>
>>
>>
>> Those raised in a morally relative or neutral environment will hold
>>             no truths to be self-evident.
>>
>> Elliott Wolin
>> Staff Physicist, Jefferson Lab
>> 12000 Jefferson Ave
>> Suite 8 MS 12A1
>> Newport News, VA 23606
>> 757-269-7365
>>
>> ================================================================================ 
>>
>>

-- 

				Sincerely,
					Elliott


================================================================================


  Those raised in a morally relative or neutral environment will hold
		    no truths to be self-evident.
				

Elliott Wolin
Staff Physicist, Jefferson Lab
12000 Jefferson Ave
Suite 8 MS 12A1
Newport News, VA 23606
757-269-7365

================================================================================