MMTx Detailed Trouble Report
(Trouble Report #)



FieldField Data
ID1
CategorymmtxT
SynopsisMissing machine output fields
Severityserious
Prioritymedium
Responsiblejim
Stateclosed
Classsw-bug
Arrival-DateThu Oct 25 09:30:00 EST 2001
Closed-DateTue Mar 19 07:11:27 EST 2002
Release---
EnvironmentOperating System: Linux
OS Version:
Java Version: Java 1.3
MySQL Version:
DescriptionActually, I have some more, but I guess they are intended: both in the candidates and the mapping lines, the second part in mmtx is missing, metamap parts such as "[[1,1],[1,1],0],[[2,2],[2,2],0]],yes,no" doe not appear in mmtx.
How-To-Repeatecho "Obstructive Sleep Apnea" | mmtx -q

NotesSee TR #13 for same problem.
Not able to duplicate here yet. I get all info.
ID2
CategorymmtxT
SynopsisMachine output differences
Severityserious
Prioritymedium
Responsibleguy
Stateclosed
Classsw-bug
Arrival-DateThu Oct 25 09:31:00 EST 2001
Closed-DateFri Mar 15 10:19:35 EST 2002
Release---
EnvironmentOperating System: Linux
OS Version:
Java Version: Java 1.3
MySQL Version:
Descriptionwith the -q option, the string "basic research" results in:
MMTx:
echo "basic research" | mmtx -a -q => utterance('00000000.tx.1',"basic research").

and MetaMap:
echo "basic research" | metamap -a -q => utterance('00000000.tx.0', "basic research").

MMTx:
echo "basic research \"a" | mmtx -a -q => utterance('00000000.tx.0', "basic research "a").

and MetaMap:
echo "basic research \"a" | metamap -a -q => utterance('00000000.tx.1',"basic research ""a")
How-To-Repeatsee above

NotesFixed in MMTx V2.0
ID3
CategorymmtxT
SynopsisResult and parsing differences
Severityserious
Prioritymedium
Responsibleguy
Stateclosed
Classsw-bug
Arrival-DateWed Oct 31 09:32:00 EST 2001
Closed-DateFri Mar 15 10:21:18 EST 2002
Release---
EnvironmentOperating System: Linux
OS Version:
Java Version: Java 1.3
MySQL Version:
Description$ echo "Raynaud's disease" | mmtx
results in:
Phrase: "Raynaud ' s disease"
Meta Candidates (10)
902 Raynaud's Disease [Disease or Syndrome]
861 Secondary disease, NOS [Disease or Syndrome]
812 Disease [Disease or Syndrome]
779 SS <1> (Sezary Syndrome) [Neoplastic Process]
779 SS <2> (Sicca Syndrome) [Disease or Syndrome]
756 DIS (Diagnostic interview schedule) [Intellectual Product]
756 Second [Quantitative Concept]
741 seconds [Temporal Concept]
719 Secondary <2> (Secondary to) [Temporal Concept]
719 secondary <1> (Metastatic to) [Functional Concept]
Meta Mapping (861)
861 Secondary disease, NOS [Disease or Syndrome]
-----------------------------------------------------------------------
$ echo "Raynaud's disease" | metamap
results in (as it should):
Phrase: "Raynaud's disease"
Meta Candidates (3):
1000 Raynaud's Disease [Disease or Syndrome]
861 Disease [Disease or Syndrome]
805 DIS (Diagnostic interview schedule) [Intellectual Product]
Meta Mapping (1000):
1000 Raynaud's Disease [Disease or Syndrome]
How-To-Repeatsee above

NotesFixed in MMTx V2.0
ID4
CategorymmtxT
SynopsisStrings not processing
Severityserious
Prioritymedium
Responsibleguy
Stateclosed
Classsw-bug
Arrival-DateSun Jan 13 09:34:00 EST 2002
Closed-DateFri Mar 15 10:22:03 EST 2002
Release---
EnvironmentOperating System: Linux
OS Version:
Java Version: Java 1.3
MySQL Version:
DescriptionFYI: the strings "beta-2 beta-3" and "n-nitro n-nitrosoguanidine" take ages to map, and actually, mmtx will die before mapping :-(
How-To-Repeatecho "beta-2 beta-3" | mmtx

NotesProblem fixed in MMTx V2.0
ID5
CategorymmtxT
SynopsisStrings not processing continued
Severityserious
Prioritymedium
Responsibleguy
Stateclosed
Classsw-bug
Arrival-DateMon Jan 14 09:35:00 EST 2002
Closed-DateFri Mar 15 10:22:51 EST 2002
Release---
EnvironmentOperating System: Linux
OS Version:
Java Version: Java 1.3
MySQL Version:
Descriptionwith regard to the two phrases I emailed yesterday. Mmtx seems to go into an infinite loop. Is it possible to let it time out if no mammping has been found?
See TR #4
How-To-Repeatecho "beta-2 beta-3" | mmtx

NotesFixed in MMTx V2.0
ID6
CategorymmtxT
SynopsisDifferences between MetaMap and MMTx
Severityserious
Prioritymedium
Responsibleguy
Statesuspended
Classsw-bug
Arrival-DateWed Mar 06 09:36:00 EST 2002
Closed-Date---
Release---
EnvironmentOperating System: Linux
OS Version:
Java Version: Java 1.3
MySQL Version:
DescriptionFYI: the sentence " the mean incidence of the 10 year period was 1.0 per 100,000 children per year."
How-To-Repeatecho "the mean incidence of the 10 year period was 1.0 per 100,000 children per year." | mmtx

NotesWorking with Lan and User Initial response is that
there is no real difference.

-----------

Looks like Version 2.2 will have shape identification.
ID7
CategorymmtxT
SynopsisRequest for modifications
Severitynon-critical
Prioritymedium
Responsiblejim
Stateclosed
Classchange-request
Arrival-DateWed Jan 30 09:39:00 EST 2002
Closed-DateWed Oct 16 11:21:44 EDT 2002
Release---
EnvironmentOperating System: Linux
OS Version:
Java Version: Java 1.3
MySQL Version:
DescriptionNOTE: Submitted for user:

- Is there an interface to the brill tagger? We do not
currently provide enough documentation (minus the source code)
to provide someone to build a tagger client.

- Is there an unistall script? Such a script would also drop
the databases.

- Is there a way to update MMTx databases short of creating
new databases and re-loading the entire models.
At issue are users that have customized the UMLS where
they are constantly changing that data to suite their
needs. User has an application database that changes
every few days.

- When making a custom UMLS, what MRSO term types get filtered out?
How-To-Repeat---
NotesSubmitted for user per email notes

1) Reference TR#16

2) Version 2.2 should have this

3) Reference TR#19

4) Will provide user with a copy of the Filtering paper.
ID8
CategorymmtxT
SynopsisInstallation Errors
Severitycritical
Priorityhigh
Responsiblejim
Stateclosed
Classsw-bug
Arrival-DateFri Feb 22 09:42:01 EST 2002
Closed-DateFri Mar 15 10:26:03 EST 2002
Release---
EnvironmentOperating System: Linux
OS Version:
Java Version: Java 1.3
MySQL Version:
DescriptionErrors during installation with respect to DB_MMTxDBTest_01_strict
How-To-Repeatinstall program

NotesFixed problems, tested, and released - MMTx V2.0.B
ID9
CategorymmtxT
SynopsisError starting up MMTx
Severitycritical
Priorityhigh
Responsiblejim
Stateclosed
Classsw-bug
Arrival-DateFri Feb 22 09:44:00 EST 2002
Closed-DateThu Jun 06 13:32:10 EDT 2002
Release---
EnvironmentOperating System: Linux
OS Version:
Java Version: Java 1.4
MySQL Version:
DescriptionI just installed MMTx, 2.0 and when I invoke it, I get the
error: "Main method is not public". How do I resolve this?
How-To-Repeatrun mmtx

NotesWork-around in V2.0.C and full fix awaiting repackaging
of MMTx. This is caused by stricter checking done in
Java 1.4
ID10
CategorymmtxT
SynopsisGeneral MMTx Question
Severitynon-critical
Prioritymedium
Responsiblejim
Stateclosed
Classquestion
Arrival-DateMon Feb 25 09:51:00 EST 2002
Closed-DateWed Oct 09 11:06:27 EDT 2002
Release---
EnvironmentOperating System: Linux
OS Version:
Java Version: Java 1.3
MySQL Version:
DescriptionCan anyone answer this? Actually, can anyone tell me how
to suppress the regular output and just get pipe delimited
output, with the options that I want? I tried
"--nmm --pipedOutput".?
How-To-RepeatN/A

NotesThis is being evaluated. We have discussed doing this in
the past - the only issue is the repetative nature of
the output.

--------------------

Looks like Version V2.2 will have this option.
ID11
CategorymmtxT
SynopsisQuestion on MMTx
Severityserious
Prioritymedium
Responsiblealan
Stateclosed
Classquestion
Arrival-DateTue Feb 26 09:52:00 EST 2002
Closed-DateTue Jul 29 14:20:57 EDT 2008
Release---
EnvironmentOperating System: Linux
OS Version:
Java Version: Java 1.3
MySQL Version:
DescriptionThe options scheme is very familiar to Unix programmers, so
I personally like it. It's also similar to the KSS API,
also familiar. What's not clear is if there are any
dependencies between the options. For example if I toggle
X on, does it have any impact on any of the other toggles,
or are they all mutually independent in their effects? I
am experimenting and it seems like some of them do have
dependencies...
How-To-RepeatN/A

NotesWe've looked at this, but no work done to date. Need to
look at MetaMap and MMTx to determine the interdependencies
of options.

------------

Same as TR #26
ID12
CategorymmtxT
SynopsisMachine output inconsistencies
Severityserious
Prioritymedium
Responsiblejim
Stateclosed
Classsw-bug
Arrival-DateTue Feb 26 09:53:00 EST 2002
Closed-DateWed Oct 16 11:20:53 EDT 2002
Release---
EnvironmentOperating System: Linux
OS Version:
Java Version: Java 1.3
MySQL Version:
Descriptionit seems if I turn off mappings or candidates, that the
output still shows a heading for them, but lists them as
zero or none, rather than just omitting mapping/candidate
information completely...

NOTE: Dependent on TR #11 & TR #26
How-To-Repeat---
NotesNeed to contact user and see what options they are using.
ID13
CategorymmtxT
SynopsisMachine output data missing
Severityserious
Prioritymedium
Responsiblejim
Stateclosed
Classsw-bug
Arrival-DateThu Feb 28 09:54:00 EST 2002
Closed-DateWed Mar 27 07:12:36 EST 2002
Release---
EnvironmentOperating System: Linux
OS Version:
Java Version: Java 1.3
MySQL Version:
Descriptionusing just the -q option, I am not getting any info after
the semantic type lists, for each candidate in a mapping
(i.e., not getting match map list, overmatch y or n, etc.). This is fine for my purposes right now, but just wondering if this is a bug?
How-To-Repeat---
NotesThis is the same as Marc's problem. Not able to duplicate
with my environment yet - I get all of the info.
ID14
CategorymmtxT
SynopsisOutput wrong
Severitycritical
Prioritymedium
Responsibleguy
Stateclosed
Classsw-bug
Arrival-DateMon Mar 04 09:55:00 EST 2002
Closed-DateWed Mar 27 07:12:53 EST 2002
Release---
EnvironmentOperating System: Linux
OS Version:
Java Version: Java 1.3
MySQL Version:
DescriptionNot sure why, but if I give the string "Lyme Disease" to
mmtx, it does not map it to the correct concept (same),
but instead maps it to "Disease".
How-To-Repeatecho "Lyme Disease" | mmtx

NotesThis has been fixed by Guy and now we are just awaiting
results from user testing.
ID15
CategorymmtxT
SynopsisNeed option to display MMTx version information
Severitynon-critical
Prioritymedium
Responsiblejim
Stateclosed
Classchange-request
Arrival-DateMon Mar 04 09:56:00 EST 2002
Closed-DateFri Mar 15 10:36:03 EST 2002
ReleaseV2.0.B
EnvironmentOperating System: Solaris
OS Version: Solaris 5.8
Java Version: Java 1.3
MySQL Version: 3.23.47
DescriptionIt would be nice to be able to type mmtx --version and receive information on
what version I'm running.
How-To-Repeat---
NotesIncorporated into MMTx V2.0.C
ID16
CategorymmtxT
SynopsisQuestion on using Brill Tagger
Severitynon-critical
Prioritymedium
Responsiblealan
Stateclosed
Classquestion
Arrival-DateFri Mar 15 20:13:00 EST 2002
Closed-DateWed May 15 11:10:02 EDT 2002
Release---
EnvironmentOperating System: Solaris
OS Version:
Java Version: Java 1.3
MySQL Version:
DescriptionNOTE: Submitted for user:

I don't see much of a problem getting the tagger up and wrapped, although
you might consider making the lisp wrapper available as well with some
documentation (I still have to figure out how to get you up from the socket
servers to CORBA-based interfaces--see our AMIA paper from a few years ago
on a laboratory test knowledge mediator using lisp<-->java connectivity
through PARC's ILU object broker). I've been looking at Brill-style taggers
over the past few months to support digital library work we are doing
in-house, of which there are also open-source distributions available.

As you point out, the tagset and the availability of tagged training texts
are significant issues. From some preliminary investigation, it seems like
there is widespread use of the treebank tagset or extensions thereof, and
there aren't health care oriented training texts available. I'm curious as
to if and how the team will be addressing this. Look forward to your
comments after the AMIA deadline.
How-To-RepeatN/A
Notes---
ID17
CategorymmtxT
SynopsisQuestion
Severityserious
Prioritymedium
Responsiblealan
Stateclosed
Classquestion
Arrival-DateThu Mar 21 09:43:00 EST 2002
Closed-DateTue Apr 02 13:28:06 EST 2002
Release---
EnvironmentOperating System: Linux
OS Version:
Java Version: Java 1.3
MySQL Version:
Description1. How can I filter/skip some obviously non-medical concepts (although they
are listed in UMLS)? For example,

(1) "Fig." inside this text should not be marked as "Fig [Food]":
This condition is called a succenturiate lobe (Fig. 6-5 ) and may be
problematic if that lobe of placenta is inadvertently left within the uterus
at the time of delivery.

(2) Likewise, "al" shouldn't be "Aluminum [Element, Ion, or Isotope]"
These studies provide strong evidence of the interconnectiveness of maternal
and fetal fluid spaces across the membranes and placenta (Kilpatrick et al,
1991).

I'd like to know a way to filter those abbreviations.
How-To-RepeatN/A
Notes---
(1) "Fig." inside this text should not be marked as "Fig [Food]"side this text should not be marked as "Fig [Food]":
This condition is called a succenturiate lobe (Fig. 6-5 ) and may is called a succenturiate lobe (Fig. 6-5 ) and may be
problematic if that lobe of placenta is inadvertently left within the uterthat lobe of placenta is inadvertently left within the uterus
at the time of deliverdelivery.
2. The other question is why MMTx includes "(" in the phrase? In the aboestion is why MMTx includes "(" in the phrase? In the above
example (2), "placenta (Kilpatrick et al" was analyzed as a phrase. Can MMlacenta (Kilpatrick et al" was analyzed as a phrase. Can MMTx
separate the---
---
I believe that a bug fix on this is being tested, but mainly because the
text wasn't broken into two phrases. As you your specific question about
the parenthesis, the parser sometimes includes punctuation that we would
normally not consider to be part of a phrase because it does not harm for
our subsequent processing. Do you have a particular reason for wanting
it excluded?
ID18
CategorymmtxT
SynopsisParsing question/problem
Severityserious
Prioritymedium
Responsibleguy
Stateclosed
Classsw-bug
Arrival-DateThu Mar 21 09:45:00 EST 2002
Closed-DateTue Jul 29 14:02:05 EDT 2008
Release---
EnvironmentOperating System: Linux
OS Version:
Java Version: Java 1.3
MySQL Version:
Description2. The other question is why MMTx includes "(" in the phrase? In the above
example (2), "placenta (Kilpatrick et al" was analyzed as a phrase. Can MMTx
separate them?

How-To-Repeatecho "placenta (Kilpatrick et al" | mmtx
echo "Nerve conduction velocity (NCV)" |mmtx
echo "Magnetic Resonance Imaging (MRI)"|mmtx
NotesThe bigger question on the fix is if it should be fixed?
If it is, it needs to be done on MetaMap as well.

This problem accounted for the bulk of the non tagger differences between MMTx and MetaMap.
ID19
CategorymmtxT
Synopsisupdate metamap dataset on-the-fly
Severityserious
Prioritymedium
Responsiblecliff
Stateclosed
Classsupport
Arrival-DateThu Mar 21 15:27:00 EST 2002
Closed-DateThu Mar 28 14:08:32 EST 2002
ReleaseV2.0B
EnvironmentOperating System: Linux
OS Version:
Java Version: Java 1.3
MySQL Version:
DescriptionI'm looking for a synopsis on how to update the metamap dataset for single concepts. Longer-term this would be a useful feature to incorporate. For the present, some documentation on what tables need to be updated would be great!
How-To-Repeat---
Notes3/22/02: In reply to TR #19:

The Data File Builder (dfbuilder) module of MMTx might be what you are
looking for. The dfbuilder was designed to allow you to create your own
data set(s) or to modify the UMLS distribution files. I would recommend that
you look at the Data File Builder documentation
(http://mmtx.nlm.nih.gov/dfBuilder.shtml) and see if that will do
what you are looking for. If it doesn't, or, if you need more information
to proceed, please provide us with more specifics on the types of updates you
are trying to do and we will work with you.


Response from our Data File Builder specialist (Cliff):

In the UMLS, information about a concept is spread over several files. So
how to modify a few concepts depends on which attributes of the concept they
want to update.

Assuming they want to add or remove troublesome synonyms they could modify
the 0Suppress1 file and follow the example of our filtering and mark them
suppressible. Subsequently building a moderate or strict model would assure the
desired terms were removed. They would start with the complete UMLS and run the
Data File Builder. If they want to try this more details can be provided.

I suggest looking at the documentation on the Data File Builder, then
sending us a more detailed description of the 'updates' they would like to make
and we may be able to suggest a process to achieve their goals.

Reply:
I will review the Data File Builder documentation.
Bill Hole should have a copy of our UMLS vocabulary server report that
details what we have done.

We support the following editing functions against umls:
create concept
Add atoms to existing concept
transpose atoms from one concept to another
delete atoms from one concept to another
merge one concept into another concept
split a concept out of an existing concept
delete concepts.

In order to integrate metamap, we need to update metamap tables
appropriately for each of these operations, as opposed to running dfbuilder
in batch mode.


Answered 4/3/02:

Changes to concepts, particularly ones that include MeSH terms, can have subtle
and possibly hard to trace consequences for MetaMap data files. For example,
splitting entry terms from a MeSH concept into separate UMLS concepts can change
not only the CUI reported for each of the words in the *words*.txt files, but
also the treecodes reported for that concept. We have not done the necessary
analysis to be able to provide a synopsis of how to update the MetaMap data
files based on changes to individual concepts.

Your study of the dfbuilder scripts and programs will probably yield the
information you seek. If you pass your findings on to us we will make them
available to other researchers on our upcoming 'contrib.' page on the MMTx web
site.
ID20
CategorymmtxT
Synopsismissing dfbuilder script 0Suppress1
Severityserious
Prioritymedium
Responsiblecliff
Stateclosed
Classquestion
Arrival-DateMon Mar 25 12:01:00 EST 2002
Closed-DateWed Mar 27 07:11:43 EST 2002
ReleaseV2.0B
EnvironmentOperating System: Linux
OS Version:
Java Version: Java 1.3
MySQL Version:
Description0Suppress1 script missing from mmtx/dfbuilder/scripts directory
How-To-Repeat---
Notes---
ID21
CategorymmtxT
SynopsisInstallation error
Severityserious
Prioritymedium
Responsiblejim
Stateclosed
Classsw-bug
Arrival-DateMon Mar 25 12:56:00 EST 2002
Closed-DateFri May 03 13:47:32 EDT 2002
Release---
EnvironmentOperating System: Linux
OS Version:
Java Version: Java 1.3
MySQL Version:
DescriptionWe are following the web guidelines required for the set up of the MMTx program. In order to run MMTx we have run ./bin/install.sh but we get this message: In Class program.MMTxinstall: main must be public and static

Would you be able to help us with this issue, namely: what does this message mean? and what, can we do in order to proceed with the installation of the software?
How-To-Repeat---
NotesLooks like we missed the installation program when we were
setting up the work-around to run in Java 1.4.

follow-up 4/2/02:

We just have one more query with regard to MMTx
installation. Indeed, we are having some trouble in
the final phase of the installation of the software.
As you may see in the output file we get when we try
to run ./bin/install.sh (please see document attached)

As referred in the output file, if the problem is the
root of Java password, please let us know how we can
introduce a new password in the root user Java.

With regard the files you sent us,after installing it,
we realised that our Java version is 1.3 so we ended
up not using the additional file you so kindly sent
us.
ID22
CategorymmtxT
SynopsisFollow-up Question #1
Severityserious
Prioritymedium
Responsiblealan
Stateclosed
Classquestion
Arrival-DateWed Mar 27 07:18:00 EST 2002
Closed-DateThu Mar 28 13:37:35 EST 2002
Release---
EnvironmentOperating System: Solaris
OS Version:
Java Version: Java 1.3
MySQL Version:
Description1. What are the 'C.......' that appear in the word files, and what is the
logic to emit them in GleanMRCON.processLine?


Answer:
C....... occurs in some data for optimization purposes. As an example, the line
abdomen|S0003328|C.......
in a word index file means that the word "abdomen" occurs in the string
S0003328 ('Palpation of abdomen'). The presence of C....... signals that
the given string is also the concept name. No further searching is necessary
to obtain the concept name.
On the other hand, the entry
abdomen|S0288461|C0000735
means that "abdomen" occurs in S0288461 ('abdomen neoplasm'), a string for
concept C0000735 with preferred name 'Abdominal Neoplasms'.
How-To-Repeat---
Notes---
ID23
CategorymmtxT
SynopsisFollow-up Question #2
Severityserious
Prioritymedium
Responsiblealan
Stateclosed
Classquestion
Arrival-DateWed Mar 27 07:19:00 EST 2002
Closed-DateThu Mar 28 13:36:47 EST 2002
Release---
EnvironmentOperating System: Solaris
OS Version:
Java Version: Java 1.3
MySQL Version:
Description2. In FilterMRCONSO.filterNmStrDups, it appears that there is no particular
criterion as to which record is kept when duplicates are discovered. Looks
like the first record processed is kept, and the dups are discarded.


Answer:
That's exactly correct; we keep the first occurrence of a string in the
mrconso file (a join of MRCON and MRSO) that is indistinguishable from
other strings within a concept. Since preferred forms occur in this file
before non-preferred forms, this has the effect of keeping the preferred
forms.
How-To-Repeat---
Notes---
ID24
CategorymmtxT
SynopsisFollow-up Question #3
Severityserious
Prioritymedium
Responsiblealan
Stateclosed
Classquestion
Arrival-DateWed Mar 27 07:19:00 EST 2002
Closed-DateThu Mar 28 13:34:54 EST 2002
Release---
EnvironmentOperating System: Solaris
OS Version:
Java Version: Java 1.3
MySQL Version:
Description3. Any particular reason for the use of another unique name
for semantic types instead of just using the unique
Tcodes?

Answer:
Yes, we find the mnemonics much easier to understand when we manually review
output.
How-To-Repeat---
Notes---
ID25
CategorymmtxT
SynopsisFollow-up Question #4
Severityserious
Prioritymedium
Responsiblealan
Stateclosed
Classquestion
Arrival-DateWed Mar 27 07:19:00 EST 2002
Closed-DateThu Mar 28 13:33:30 EST 2002
Release---
EnvironmentOperating System: Solaris
OS Version:
Java Version: Java 1.3
MySQL Version:
Description4. Why isn't jlvg used for normalization? NLSStrings looks to be more
lightweight, was it just easier to include?

Answer:
NLSStrings is a descendent of a corresponding Prolog module used by MetaMap
and similar programs developed in the Natural Language Systems program.
It predates lvg, is tailored to our specific needs, and is more efficient
than lvg (the last time I checked). I'm also fairly certain that lvg doesn't
have the full functionality required for the specific kind of normalization
we're doing; in particular, it doesn't respect word order and the kinds of
normalization involving NOS and NEC are different (MetaMap continues to
filter out NOS but no longer does so with NEC).
How-To-Repeat---
Notes---
ID26
CategorymmtxT
SynopsisExplanation of options with interdependencies
Severityserious
Prioritymedium
Responsiblealan
Stateclosed
Classdoc-bug
Arrival-DateWed Mar 27 09:05:01 EST 2002
Closed-DateTue Jul 29 14:21:20 EDT 2008
Release2.0.C
EnvironmentOperating System: Solaris
OS Version: 2.8
Java Version: Java 1.2
MySQL Version: 3.23.39
DescriptionThe documentation needs to include adequate option descriptions
along with explanations of how they interact.

----------

Same as TR#11
How-To-Repeat---
Notes---
ID27
CategorymmtxT
Synopsis-h needs to be translated into English
Severityserious
Prioritymedium
Responsiblejim
Stateclosed
Classdoc-bug
Arrival-DateWed Mar 27 09:08:00 EST 2002
Closed-DateTue Jun 04 14:14:33 EDT 2002
Release2.0.C
EnvironmentOperating System: Solaris
OS Version: 2.8
Java Version: Java 1.2
MySQL Version: 3.23.39
DescriptionThe output from -h is unreadable. It needs to be written so that options
are organized into groups (see the web site), so that columns are straight,
and so that debug options are somehow isolated (perhaps with a new option?).
How-To-Repeatmmtx -h
NotesThis has been done in a sample file and made available
to the user for comment. All we are waiting to do at this
point is to install into the MMTx code.
ID28
CategorymmtxT
SynopsisMMTx hangs when running the following
Severityserious
Prioritymedium
Responsibleguy
Stateclosed
Classsw-bug
Arrival-DateFri Mar 29 06:34:00 EST 2002
Closed-DateWed May 15 11:12:23 EDT 2002
Release---
EnvironmentOperating System: Solaris
OS Version:
Java Version: Java 1.3
MySQL Version:
DescriptionExecuting mmtx on the following text:

Weight Reduction^Excess body weight-body mass index
causes it to hang indefinitely...... Thought I should let you know.
I don't think we have yet installed any of the bug fixes for v 2.0 on our
machines....maybe they would fix this, I don't know.

-----------------

Output differs greatly from MetaMap
How-To-Repeatrun text through MMTx
NotesI believe it works in the current release 2.0.C. I get
the proper answer when I run it through - please see if you
can update to the 2.0.C version and run it and see if you
are getting the same results. I've tested this query in
both V2.0.B and V2.0.C and get the same results. Are you
using any options? I only tried it with the default options
set.

-----------------------------
ID29
CategorymmtxT
SynopsisMMTx doesn't work after applying updates for V2.0.C
Severityserious
Priorityhigh
Responsiblejim
Stateclosed
Classsw-bug
Arrival-DateMon Apr 01 06:48:01 EST 2002
Closed-DateWed Apr 03 13:14:10 EST 2002
Release---
EnvironmentOperating System: Solaris
OS Version:
Java Version: Java 1.3
MySQL Version:
DescriptionWe installed patches to version 2.0.C today.

I just tried to run the text again and now it doesn't hang, but gives the
following java error:

Using Database: DB_01_strict
Weight Reduction^Excess body weight-body mass index
^D

Processing 00000000.tx.0: Weight Reduction^Excess body weight-body mass index
Exception in thread "main" java.lang.NoSuchMethodError
at
lexicalization.LexicalLookup.makeLexicalElementOf(LexicalLookup.java:1307)
at
lexicalization.LexicalLookup.matchMultiWords(LexicalLookup.java:941)
at lexicalization.LexicalLookup.lexicalLookup(LexicalLookup.java:331)
at
lexicalization.LexicalLookup.lexicalLookupAux(LexicalLookup.java:239)
at
lexicalization.LexicalLookup.lexicalLookupAux(LexicalLookup.java:212)
at programs.MMTx.processSentence(MMTx.java:310)
at programs.MMTx.processDocumentX(MMTx.java:1161)
at programs.MMTx.processCollectionX(MMTx.java:876)
at programs.MMTx.run(MMTx.java:1049)
at programs.MMTx.main(MMTx.java:1235)
How-To-Repeat---
Notes Thanks for the quick return on the info. It looks like you are
running a very early version of MMTx. In reference to your
question on how to tell which version - unfortunately, until V2.0.C
there really isn't except to look at the dates of the files.
In looking at the dates of the class files from the listing
provided. Most of your classes are from the November 2001 timeframe
which was the timeframe when we had the early V1.0.? where ? is one
of A-I. I'm "guessing" from the date, that it is V1.0.I since that
would have been the latest version available in late-November.

What I'd like to do I guess is build a tar file to replace all of
your existing files. This would have the benefit of not forcing you
to go back through the installation process. I just need to know if
you changed any of the source files before I setup the tar file for
you.

Sorry for all the confusion, but, that does explain why I'm not able
to duplicate some of the problems you were seeing.

PS: We will also need to work on the script that you are using to
run MMTx with. The updated files should have a new one ready for you.
ID30
CategorymmtxT
SynopsisException occurred when parsing a sentence
Severityserious
Priorityhigh
Responsiblejim
Stateclosed
Classsw-bug
Arrival-DateWed Apr 03 10:34:00 EST 2002
Closed-DateMon Apr 08 07:19:56 EDT 2002
ReleaseV2.0
EnvironmentOperating System: Windows NT/2000
OS Version:
Java Version: Java 1.3
MySQL Version:
DescriptionWhen parsing this sentence:

In a joint U.S. Centers for Disease Control and Prevention and Thai Ministryof Public Health trial, 397 asymptomatic Thai women were randomized to receive either placebo or AZT in the last month of pregnancy(300 mg tablets orallytwice dailyfrom 36 weeks until labor, one tablet at the onset of labor, and one tablet every3 hours until delivery).

MMTx generated this error:


Something went wrong: java.lang.ArrayIndexOutOfBoundsException
java.lang.ArrayIndexOutOfBoundsException
at mm.Candidate.doCandidatePosCheck(Candidate.java:797)
at candidateGeneration.CandidateGeneration.computeMetaCandidates(Candida
teGeneration.java:484)
at candidateGeneration.CandidateGeneration.createCandidates(CandidateGen
eration.java:222)
at programs.MMTx.mapPhrase(MMTx.java:469)
at programs.MMTx.processPhrase(MMTx.java:412)
at programs.MMTx.processSentence(MMTx.java:347)
at programs.MMTx.processDocumentX(MMTx.java:1154)
at programs.MMTx.processCollectionX(MMTx.java:874)
at programs.MMTx.run(MMTx.java:1041)
at programs.MMTx.main(MMTx.java:1228)



04/05/02:

I have downloaded V2.0C and tried the same test, but still failed. It seemed
that my original 02/15/2002 version is V2.0A. I upgraded from V2.0A directly
to V2.0C. This should be fine, right? Since V2.0B it's only a fix for
database installation process.
Anyway, I have listed all files under my mmtx/classes directory. Attached
please find this file. For some newer class files, it might because I
recompiled some source code. But I didn't break any original functionality.
For that mm/Candidate.class, which caused the exception, I only changed one
line in the source code so that my test could be passed.

Answer:

When you say "I have downloaded V2.0C" and upgraded over V2.0A what
exactly do you mean? Did you download the full V2.0C distribution
and reinstall? Or did you download the upgrade for V2.0C and upgrade?

Also, the changing one line raises a yellow flag on my side. Could you
forward me the source code of the file you changed and I'll try running
with that file to see if I can replicate your problem.


04/05/02:

I downloaded only the upgrade files, i.e., "Version 2.0.C of MMTx
Classes and Source Update (85kb)", then did the upgrade.
Attached please find Candidate.java. You can search "Huang" for the line I
changed in order to pass my test. I didn't have enough time to dig into more
deeply into the code.


Answer:

The upgrade was only designed to take you from V2.0B to V2.0C and
only included those files.

I've compiled and ran with your changed file and your right it isn't
changing the behavior. It runs fine on my system and without the
error. Under both Solaris and Windows 2000.

I believe the error is coming in because of the old version mixed with
the new version. The only two solutions I can offer are the following:

1) Do a complete reinstall with the V2.0.C source & classes distribution.

2) Download the complete V2.0.C source & classes distribution, tar the
contents into a new directory and copy over the source files and
class files into your existing mmtx directory.
How-To-Repeat---
NotesInitially, I'm not able to reproduce on the Solaris
platform with V2.0.B or V2.0.C. Need to check on the
PC side to see if it is platform specific.

4/4/02: Unable to reproduce on Windows platform.

------------------

I'm in the process of reloading my PC with MMTx and then I will try
running your file through. I don't get the same error on the Sun
Solaris platform, so, I need to check to see if it is something
specific to the PC.

If you are running from a version on 2/15, you might want to consider
upgrading if any of the changes outlined on the download page apply
to the kinds of work you are doing.

I'll let you know when I've had a chance to run on the PC side.
ID31
CategorymmtxT
SynopsisfullVars from GenerateMMVariants does not match MetaMap tables
Severityserious
Prioritymedium
Responsibleguy
Statesuspended
Classsw-bug
Arrival-DateWed Apr 03 15:58:00 EST 2002
Closed-Date---
Release2.0.A
EnvironmentOperating System: Solaris
OS Version: Solris 2.8
Java Version: Java 1.3
MySQL Version:
Description I have attempted to validate the variant generation that is part of the
MMTx Data File Builder. There are differences for which I do not know the
significance. Here are my observations, please let me know if they are
significant.

The full01.txt is distributed with MMTx and is a join of the variant tables
used by MetaMap. The fullVars_02_custom.txt was produced by GenerateMMVariants
on a Linux box. I have sorted them, removed the final pipe from the former and
compared the resulting files. Those files can be found in
/home/cliff/projects/data_file_builder/variants
The files used were
full01.sorted.stripped -- Herein known as Lan's
fullVars_02_custom.sorted -- Herein known as Guy's

1) Size: Lan's 1 592 620
Guy's 1 812 481

2) Common Records: 8631

3) Numbers: Lan's about 207 000
Guy's about 94 000
Lan set the second column of numbers to none, guy set column two to the
value in column 4.

4) None: Lan used it 982 021 times
Guy never used it.

5) Example 1:

I picked a word at random where I found some common rows and compared the
treatement from both files. The indented values come from Guy's file. The plus
indicated a record unique to that file.

Lan's version does not have noun to verb transitions.
Guy's version has some duplicate records, that may not pose a problem in
the database.
There are some differences even for similar records.

absence|noun|absence|noun|0|NULL|1|1|1|1|
absence|noun|absences|noun|1|i|1|1|1|1|
absence|noun|absent|adj|3|d|1|1|1|1|

absence|noun|absence|noun|0|NULL|1|1|1|1
absence|noun|absences|noun|1|i|1|1|1|1
absence|noun|absent|adj|3|d|1|1|1|1
+absence|noun|absented|verb|7|ddi|1|0|0|1
+absence|noun|absenting|verb|7|ddi|1|0|0|1
+absence|noun|absents|verb|7|ddi|1|0|0|1
+absence|noun|absent|verb|6|dd|1|0|0|1


absent|adj|absence|noun|3|d|1|1|1|1|
absent|adj|absences|noun|4|id|1|1|1|1|
absent|adj|absent|adj|0|NULL|1|1|1|1|
absented|verb|absents|verb|1|i|1|1|1|1| (a)
absented|verb|absent|verb|1|i|1|1|1|1|

absent|adj|absence|noun|3|d|1|1|1|1
absent|adj|absences|noun|4|di|1|1|1|1
absent|adj|absent|adj|0|NULL|1|1|1|1
+absented|verb|absented|verb|0|NULL|1|1|1|1
+absented|verb|absented|verb|0|NULL|1|1|1|1 --duplicate
+absented|verb|absenting|verb|2|ii|1|1|1|1
absented|verb|absents|verb|2|ii|1|1|1|1 differs from (a)
absented|verb|absent|verb|1|i|1|1|1|1


6) Example 2:

The recording of the history is reversed for these to files.
Guy's has some new variants, are they valid?

ocular|adj|eye|noun|2|s|1|1|1|1
ocular|adj|eyes|noun|3|is|1|1|1|1
ocular|adj|ocular|adj|0|NULL|1|1|1|1
ocular|adj|oculi|noun|4|id|1|1|1|1
ocular|adj|oculus|noun|3|d|1|1|1|1
ocular|adj|ophthalmia|noun|7|dss|1|1|1|1
ocular|adj|ophthalmic|adj|4|ss|1|1|1|1
ocular|adj|optic|adj|4|ss|1|1|1|1

ocular|adj|eye|noun|2|s|1|1|1|1
ocular|adj|eyes|noun|3|si|1|1|1|1
ocular|adj|ocular|adj|0|NULL|1|1|1|1
ocular|adj|oculi|noun|4|di|1|1|1|1
ocular|adj|oculus|noun|3|d|1|1|1|1
+ocular|adj|ophthalmiac|noun|7|ssd|1|1|1|1
+ocular|adj|ophthalmiacs|noun|8|ssdi|1|1|1|1
ocular|adj|ophthalmia|noun|7|ssd|1|1|1|1
+ocular|adj|ophthalmias|noun|8|ssdi|1|1|1|1
ocular|adj|ophthalmic|adj|4|ss|1|1|1|1
ocular|adj|optic|adj|4|ss|1|1|1|1


eye|noun|eye|noun|0|NULL|1|1|1|1
eye|noun|eyes|noun|1|i|1|1|1|1
eye|noun|ocular|adj|2|s|1|1|1|1
eye|noun|oculi|noun|6|ids|1|1|1|1
eye|noun|oculus|noun|5|ds|1|1|1|1
eye|noun|ophthalmia|noun|5|ds|1|1|1|1
eye|noun|ophthalmic|adj|2|s|1|1|1|1
eye|noun|optic|adj|2|s|1|1|1|1

eye|noun|eye|noun|0|NULL|1|1|1|1
eye|noun|eyes|noun|1|i|1|1|1|1
eye|noun|ocular|adj|2|s|1|1|1|1
eye|noun|oculi|noun|6|sdi|1|1|1|1
eye|noun|oculus|noun|5|sd|1|1|1|1
+eye|noun|ophthalmiac|noun|5|sd|1|1|1|1
+eye|noun|ophthalmiacs|noun|6|sdi|1|1|1|1
eye|noun|ophthalmia|noun|5|sd|1|1|1|1
+eye|noun|ophthalmias|noun|6|sdi|1|1|1|1
eye|noun|ophthalmic|adj|2|s|1|1|1|1
eye|noun|optic|adj|2|s|1|1|1|1

7) Example 3:

I picked out a number to show differences.

For numbers Lan has second column with none, and an extra row with none in
column four:
105058|none|105058|adj|0|NULL|1|1|1|1
105058|none|105058|adv|0|NULL|1|1|1|1
+105058|none|105058|none|0|NULL|1|1|1|1
105058|none|105058|noun|0|NULL|1|1|1|1
105058|none|105058's|noun|1|i|1|1|1|1
105058|none|105058|verb|0|NULL|1|1|1|1
105058|adj|105058|adj|0|NULL|1|1|1|1
105058|adv|105058|adv|0|NULL|1|1|1|1
105058|noun|105058|noun|0|NULL|1|1|1|1
105058|noun|105058's|noun|1|i|1|1|1|1
105058|verb|105058|verb|0|NULL|1|1|1|1

99994|none|99994|adj|0|NULL|1|1|1|1
99994|none|99994|adv|0|NULL|1|1|1|1
+99994|none|99994|none|0|NULL|1|1|1|1
99994|none|99994|noun|0|NULL|1|1|1|1
99994|none|99994's|noun|1|i|1|1|1|1
99994|none|99994|verb|0|NULL|1|1|1|1

99994|adj|99994|adj|0|NULL|1|1|1|1
99994|adv|99994|adv|0|NULL|1|1|1|1
99994|noun|99994|noun|0|NULL|1|1|1|1
99994|noun|99994's|noun|1|i|1|1|1|1
99994|verb|99994|verb|0|NULL|1|1|1|1

8) Example 4:

This abbreviation showed some similarities, but the treatment seemed to be
substantially different.

Lans:
abo's|noun|abo|noun|1|i|1|1|1|1
abos|noun|abo|noun|1|i|1|1|1|1
abo's|noun|aborted|verb|6|ide|1|0|0|0
abos|noun|aborted|verb|6|ide|1|0|0|0
abo's|noun|aborter|noun|8|dde|1|0|0|0
abos|noun|aborter|noun|8|dde|1|0|0|0
abo's|noun|abortion|noun|2|e|1|1|1|0
abos|noun|abortion|noun|2|e|1|1|1|0
abo's|noun|abortions|noun|3|ie|1|1|1|0
abos|noun|abortions|noun|3|ie|1|1|1|0
abo's|noun|abort|verb|5|de|1|0|0|0
abos|noun|abort|verb|5|de|1|0|0|0
abo's|noun|abo's|noun|0|NULL|1|1|1|1
abos|noun|abo's|noun|0|p|1|1|1|1
abo's|noun|absent bed occupancy|noun|2|e|1|1|1|0
abos|noun|absent bed occupancy|noun|2|e|1|1|1|0
abo's|noun|absent bed occupant|noun|2|e|1|1|1|0
abos|noun|absent bed occupant|noun|2|e|1|1|1|0
abo's|noun|antibodies|noun|3|ie|1|1|1|0
abos|noun|antibodies|noun|3|ie|1|1|1|0
abo's|noun|antibody|noun|2|e|1|1|1|0
abos|noun|antibody|noun|2|e|1|1|1|0
abo's|noun|preventive substance|noun|4|se|1|1|1|0
abos|noun|preventive substance|noun|4|se|1|1|1|0
abo's|noun|sensitizing substance|noun|4|se|1|1|1|0
abos|noun|sensitizing substance|noun|4|se|1|1|1|0
abo's|noun|substance sensibilisatrice|noun|4|se|1|1|1|0
abos|noun|substance sensibilisatrice|noun|4|se|1|1|1|0


Guy's:
ABO's|noun|ABO|noun|1|i|1|1|1|1
ABOs|noun|ABO|noun|1|i|1|1|1|1
ABO's|noun|aborted|verb|7|iadi|1|0|0|0
ABOs|noun|aborted|verb|7|iadi|1|0|0|0
ABO's|noun|aborter|noun|9|iadd|1|0|0|0
ABOs|noun|aborter|noun|9|iadd|1|0|0|0
ABO's|noun|aborters|noun|10|iaddi|1|0|0|0
ABOs|noun|aborters|noun|10|iaddi|1|0|0|0
ABO's|noun|aborting|verb|7|iadi|1|0|0|0
ABOs|noun|aborting|verb|7|iadi|1|0|0|0
ABO's|noun|abortionist|noun|6|iad|1|0|0|0
ABOs|noun|abortionist|noun|6|iad|1|0|0|0
ABO's|noun|abortionists|noun|7|iadi|1|0|0|0
ABOs|noun|abortionists|noun|7|iadi|1|0|0|0
ABO's|noun|abortion|noun|3|ia|1|1|0|0
ABOs|noun|abortion|noun|3|ia|1|1|0|0
ABO's|noun|abortions|noun|4|iai|1|1|0|0
ABOs|noun|abortions|noun|4|iai|1|1|0|0
ABO's|noun|aborts|verb|7|iadi|1|0|0|0
ABOs|noun|aborts|verb|7|iadi|1|0|0|0
ABO's|noun|abort|verb|6|iad|1|0|0|0
ABOs|noun|abort|verb|6|iad|1|0|0|0
ABO's|noun|ABO's|noun|0|NULL|1|1|1|1
ABOs|noun|ABOs|noun|0|NULL|1|1|1|1
ABO's|noun|ABOs|noun|2|ii|1|1|1|1
ABOs|noun|ABO's|noun|2|ii|1|1|1|1
ABO's|noun|antibodies|noun|4|iai|1|1|0|0
ABOs|noun|antibodies|noun|4|iai|1|1|0|0
ABO's|noun|antibody|noun|3|ia|1|1|0|0
ABOs|noun|antibody|noun|3|ia|1|1|0|0
ABO's|noun|ectrotic|adj|5|ias|1|1|0|0
ABOs|noun|ectrotic|adj|5|ias|1|1|0|0
How-To-Repeat---
Notes---
ID32
CategorymmtxT
Synopsisdifferent output for "activated" between mmtx and skr
Severityserious
Prioritymedium
Responsibleguy
Stateclosed
Classsw-bug
Arrival-DateTue Apr 30 06:31:00 EDT 2002
Closed-DateTue Jul 29 14:02:23 EDT 2008
ReleaseV2.0.C
EnvironmentOperating System: Linux
OS Version: redhat 7.0. problem also on OS X
Java Version: Java 1.3
MySQL Version: 3.23.49a
Descriptionskr and mmtx differ in mapping between "activated". mmtx
does not find any mapping, skr finds "Activate" as concept.
How-To-Repeatecho "activate" | MMTx -a -q
Notes---
ID33
CategorymmtxT
Synopsismachine output not consistent, related to verb inflections
Severitycritical
Priorityhigh
Responsiblejim
Stateclosed
Classsw-bug
Arrival-DateTue Apr 30 06:34:00 EDT 2002
Closed-DateThu May 30 09:06:45 EDT 2002
ReleaseV2.0.C
EnvironmentOperating System: Linux
OS Version: redhat 7.0. problem also on OS X
Java Version: Java 1.3
MySQL Version: 3.23.49a
Descriptionmachine output is not always consistent when there is no mapping found.
echo "activate" | MMTx -a -q results in:

utterance('00000000.tx.1',"activated").
phrase('activated',[verb([lexmatch([activated]),inputmatch([activated]),tag(verb),tokens([activated])])]).
candidates([]).mappings([]).
'EOU'.

the mappings ios behind the candidates, and not on a new line!!. same problem with talking, signaling, deleted, disrupts, oxidized, i.e. some verb inflections lead to this problematic output
How-To-Repeatecho "activate" | MMTx -a -q
NotesProblem in MMTx.java where newline wasn't being placed
after zero candidates list found. Should be in new
release 2.0.D
ID34
CategorymmtxT
Synopsismmtx / skr inconsistency, related to hyphens?
Severityserious
Prioritymedium
Responsibleguy
Stateclosed
Classsw-bug
Arrival-DateTue Apr 30 07:28:01 EDT 2002
Closed-DateThu Apr 24 12:12:40 EDT 2003
ReleaseV2.0.C
EnvironmentOperating System: Linux
OS Version: redhat 7.0. problem also on OS X
Java Version: Java 1.3
MySQL Version: 3.23.49a
Descriptiontake "interleukin-10" and compare mmtx and skr: SKR finds "Interleukin-10", but SKR only finds "Interleukins, NOS". In machine output: it seems that MMTx inserts a space in front of the hyphen


MMTx output after MMTx -a -q:
utterance('00000000.tx.1',"interleukin-10").
phrase('interleukin -10',[head([lexmatch([interleukin]),inputmatch([interleukin]),tag(noun),tokens([interleukin])]),not_in_lex([inputmatch([-10]),tag(unknown),tokens([-,10])])]).
candidates([ev(-827,'Interleukin, NOS','Interleukins',[interleukin],[aapp,imft],[[[1,1],[1,1],0]],yes,no),ev(-755,'ILS','ILS',[ils],[aapp,phsu],[[[1,1],[1,1],3]],yes,no)]).
mappings([map(-827,[ev(-827,'Interleukin, NOS','Interleukins',[interleukin],[aapp,imft],[[[1,1],[1,1],0]],yes,no)])]).
'EOU'.



SKR output:
utterance('00000000.tx.1',"interleukin-10").
phrase('interleukin-10',[mod([lexmatch([interleukin]),inputmatch([interleukin]),tag(noun),tokens([interleukin])]),punc([inputmatch([-]),tokens([])]),shapes([inputmatch(['10']),features([integer]),tokens(['10'])])]).
candidates([ev(-1000,'Interleukin-10','Interleukin-10',[interleukin,'10'],[aapp,imft],[[[1,1],[1,1],0],[[2,2],[2,2],0]],yes,no),ev(-861,'Interleukin, NOS','Interleukins',[interleukin],[aapp,imft],[[[1,1],[1,1],0]],yes,no)]).
mappings([map(-1000,[ev(-1000,'Interleukin-10','Interleukin-10',[interleukin,'10'],[aapp,imft],[[[1,1],[1,1],0],[[2,2],[2,2],0]],yes,no)])]).
'EOU'.
How-To-Repeatecho "activate" | MMTx -a -q
Notes---
ID35
CategorymmtxT
Synopsiserror on file generation when creating custom knowledge source
Severityserious
Priorityhigh
Responsiblecliff
Stateclosed
Classsw-bug
Arrival-DateFri May 17 14:38:00 EDT 2002
Closed-DateThu Jun 06 13:30:52 EDT 2002
Release2.0D
EnvironmentOperating System: Windows NT/2000
OS Version: win2k
Java Version: Java 1.4
MySQL Version: 3.23.42
DescriptionProblem 1:

1)Running 05GenerateMWIfiles, always get 0 size file for cui_concept, cui.concept.st and concept_st.txt
2)the first line of file "cui_concept" is strange, like this:
C.......|X
C0000005|A1BG
C0000005|alpha-1-B glycoprotein
C0000007|A2M
C0000007|alpha-2-macroglobulin

2) Running LoadDataFiles, can not load table "concept_cui" and "concept_st", check the file "concept_cui.txt.0" and "concept_cui.txt" having strange format:

X
|C.......
A1BG
|C0000005
alpha-1-B glycoprotein
|C0000005
A2M
|C0000007
alpha-2-macroglobulin
|C0000007
A2MP
|C0000008

It seems there is a extra "\n" after the first field.

How-To-RepeatRun all procedures to 05GenerateMWIFiles
Run LoadDataFiles
NotesAdd special condition to filter out
extraneous ^Ms from java output files.
ID36
CategorymmtxT
SynopsisBuildDataFiles fails because of bad path
Severitycritical
Priorityhigh
Responsiblewill
Stateclosed
Classsw-bug
Arrival-DateFri May 31 11:08:01 EDT 2002
Closed-DateWed Oct 09 10:59:55 EDT 2002
Release2.0.D
EnvironmentOperating System: Windows NT/2000
OS Version:
Java Version: Java 1.3
MySQL Version:
DescriptionIf the installation of MMTx in a directory below the top level of a device,
the path offered for the creation of the workspace is wrong. Instead of
/cygwin/e/MYTOP/mmtx/... you get E:MYTOPmmtx/...
This also affects LoadDataFiles and the dfbuilder.profile.
These all need double back slashes \\ replacement by the
setup script because they are going to be interpreted in a
shell environment that will consume single back slashes.
How-To-Repeat---
Notes---
ID37
CategorymmtxT
SynopsisTest of email alias
Severityserious
Prioritymedium
Responsiblejim
Stateclosed
Classsw-bug
Arrival-DateMon Jun 03 07:10:00 EDT 2002
Closed-DateMon Oct 07 09:59:28 EDT 2002
ReleaseV2.0.D
EnvironmentOperating System: Solaris
OS Version: Solaris 2.8
Java Version: Java 1.3
MySQL Version: 3.23.49
DescriptionJust wanted to test to see if Lan receives this email via mail alias
on SKR1.
How-To-Repeat---
Notes---
ID38
CategorymmtxT
Synopsisinstallation scripts gives error
Severitycritical
Prioritymedium
Responsiblewill
Stateclosed
Classsw-bug
Arrival-DateWed Jun 05 10:37:00 EDT 2002
Closed-DateThu Jun 06 13:25:41 EDT 2002
ReleaseV2.0.D
EnvironmentOperating System: MacOS-X
OS Version: 10.1.5
Java Version: Java 1.3
MySQL Version: 3.23.46
Descriptioninstallation script gives error:

JGM: Database mmtx does not exist - creating ... ERROR
-- Unable to create Database mmtx exitting application


Root access is available, in MySQL, the strict, mmtx and lexicon databases have been created. However, no tables are in those databases.

Log file is:
-----------------------------------------------------------

Welcome to the MMTx Installation script

We will be walking you through the following:
0) Validating environment,
1) Setting up the pathing in the config directory files,
2) Unloading the data files,
3) Setting up the mmtxUser MySQL user,
4) Creating necessary databases in MySQL,
5) Loading necessary tables in MySQL and,
6) Running an initial test to verify operations.


Note: The installation script uses some GNU CYGWIN programs that
have been distributed along with this script. These commands
are only used when this installation is on a Windows platform.
The command that have been distributed include tar, gunzip,
rm, find, hosthame, and cmp.

The CYGWIN package is licensed under the GNU Public License.
The entire CYGWIN package, along with the sources to these
commands can be found at http://sources.redhat.com/cygwin/.
-----------------------------------------------------------

NOTE: This script requires root privilege in MySQL!
Continue[Y/y or N/n] [N]? File Found :/usr/bin/tar
Is this the correct instantiation to use? [Y/y or N/n] [Y]? File Found :/usr/bin/gunzip
Is this the correct instantiation to use? [Y/y or N/n] [Y]? File Found :/bin/rm
Is this the correct instantiation to use? [Y/y or N/n] [Y]? File Found :/usr/bin/cmp
Is this the correct instantiation to use? [Y/y or N/n] [Y]?
1) Setting up the pathing in the config directory files:

Full pathname of the mmtx PARENT directory [/Users/weeber/MMTx]: Specify the Database Server Host name that you will be using.
The default is the database engine or server (mysql or some other db) will
be on the same machine that the MMTx sofware is on.

Database HostName [localhost]:
Running Setup ...
-- Updating hostname in config/mmtxRegistry.cfg ...
Done
-- Updating hostname in resources/mysql.cmd ...
Done

2) Unloading the data files:
Done

3) Setting up the mmtxUser MySQL user:

Change the mmtxUser default password (Y/y or N/n) [N]? -- You have elected to use the default password: 'defaultPass'

3 & 4) Finish setting up mmtxUser and creating Databases:

Database root user[root]:
Does the database have a password set for user internal (Y/y or N/n) [N]? Enter the Mysql Root password

The following MMTx elements are already installed:
-- Nothing currently installed

-- Making sure we have a clean MySQL first ...
Done

-- Adding user and tables to MySQL ...
Done


5) Loading necessary tables in MySQL:

PLEASE NOTE: This step takes some time due to the large
amount of data loaded and the creation of indexes.
We will be automatically going through all of the databases
you downloaded. The estimates are based on a Sun Sparc Ultra 60.
The Windows estimates are based on a 400 MHZ PENTIUM II, running NT2000.

-- mmtx ~1 hour and 40 minutes
-- Strict Model (default) ~55 minutes [45 minutes on Windows]
-- Moderate Model (optional) ~1 hour and 40 minutes
-- Relaxed Model (optional) ~2 hours
-- Lexicon ~20 minutes [42 minutes on Windows

-- Loading mmtx tables ...
-- Creating the Variants tables and the Semantic Types table...
How-To-Repeat---
NotesStuffit expander incorrectly extracts GNU tar archives.
ID39
CategorymmtxT
SynopsisData file builder when generating variant files
Severityserious
Prioritymedium
Responsibleguy
Stateclosed
Classsw-bug
Arrival-DateFri Jun 07 04:27:00 EDT 2002
Closed-DateThu Apr 24 12:14:00 EDT 2003
ReleaseV2.0.D + patch from Willy
EnvironmentOperating System: MacOS-X
OS Version: 10.1.5
Java Version: Java 1.3
MySQL Version: 3.23.46
Descriptionwhen usiong the data file builder to extract only MESH from the UMLS, the generation of variant files (executing 01GenerateVariants in the directory 03variants) results in an error. Note that the java programs are not killed. Output is written in the directory subInputDir. The contens of those files look like:


......
bikunin|-2147483632
curcacycline A|-2147483632
didimethylsulfoxide|-2147483632
spoliation|-2147483632
.....


The error message is:


** Configuration Error: Can't find bundle for base name data.config.lvg, locale en_US
** Error: problem of opening/reading config file: 'data.config.lvg'. Use -x option to specify the config file path.
java.lang.NullPointerException
at Lvg.Lib.Configuration.GetConfiguration(Configuration.java:59)
at Lvg.Api.LvgApi.Init(LvgApi.java:225)
at Lvg.Api.LvgApi.(LvgApi.java:62)
at variantGeneration.LexicalTools.(LexicalTools.java:49)
at variantGeneration.PGvSlave.generateVariants(PGvSlave.java:308)
at variantGeneration.PGvSlave.generateVariants(PGvSlave.java:213)
at variantGeneration.PGvWorker.run(PGvWorker.java:100)
at variantGeneration.PGvWorker.main(PGvWorker.java:330)
java.lang.NullPointerException
at Lvg.Lib.Configuration.GetConfiguration(Configuration.java:59)
at Lvg.Api.LvgApi.Init(LvgApi.java:225)
at Lvg.Api.LvgApi.(LvgApi.java:62)
at variantGeneration.LexicalTools.(LexicalTools.java:49)
at variantGeneration.PGvSlave.generateVariants(PGvSlave.java:308)
at variantGeneration.PGvSlave.generateVariants(PGvSlave.java:213)
at variantGeneration.PGvWorker.run(PGvWorker.java:100)
at variantGeneration.PGvWorker.main(PGvWorker.java:330)



How-To-Repeat---
Notes---
ID42
CategorymmtxT
SynopsisFruitful data not in proper format.
Severitynon-critical
Prioritymedium
Responsibleguy
Stateclosed
Classsw-bug
Arrival-DateFri Aug 09 12:31:00 EDT 2002
Closed-DateThu Apr 24 12:19:13 EDT 2003
Release2.2.A
EnvironmentOperating System: MMTx: Solaris MySQL:Solaris
OS Version: 2.8
Java Version: Java 1.4
MySQL Version:
DescriptionThe lvg installations yields an error with the checked in data.
Fruitful.data yields the following message from MySQL:
-- Loading data into Fruitful table ...
Exception: java.lang.NumberFormatException: n+s
Line: 0 - !kung|!Kung|128|1|Gn|1|n+s|0|
***Error: problem of reading file.

-- Total Inserted Record: 0

Guy told us that this did not matter for Variant Generation.
The file has been replaced with an empty file which generates no errors.

Valid data should be checked in, or this problem can be closed if the data
is really not needed.
How-To-RepeatRun mmtx install and choose to install the data file builder.
Notes---
ID43
CategorymmtxT
SynopsisMMTx fails verification test during install, but otherwise runs normally
Severityserious
Prioritymedium
Responsiblewill
Stateclosed
Classsw-bug
Arrival-DateMon Aug 19 14:35:00 EDT 2002
Closed-DateWed Oct 09 10:59:26 EDT 2002
ReleaseMMTx V2.2.A
EnvironmentAffects Solaris 2.8, Red Hat Linux 7.2, Mac OS/X 10.1, Windows NT, and possibly others.
DescriptionMMTx fails verification test during install but runs correctly
afterwards from command line.
How-To-Repeatrun sh ./bin/install.sh on UNIX or ./bin/install.bat on
windows from MMTx directory.
NotesRemoved leading "./" from "config" in CLASSPATH:

CLASSPATH=./classes:./config:./lib/mmtx.jar: ...

becomes:

CLASSPATH=classes:config:./lib/mmtx.jar: ...

However, the real fix is to use ClassLoader.getSystemResourceAsStream(String) instead.
I modified GlobalBehavior.readRegistryFile(String) and
GobalBehavior.readConfigFile(String) to use getSystemResourceAsStream(String).
ID45
CategorymmtxT
SynopsisErrors on database actions not reported.
Severityserious
Prioritymedium
Responsiblewill
Stateclosed
Classchange-request
Arrival-DateFri Aug 30 10:30:01 EDT 2002
Closed-DateWed Oct 09 13:33:05 EDT 2002
Release---
EnvironmentOperating System: MMTx: Solaris MySQL:Solaris
OS Version: 2.8
Java Version: Java 1.4
MySQL Version:
DescriptionEncountered this problem when error in building the paths for datafiles on the
PC resulted in verification errors at the end of the installation.
The DbUtil class sends SQL commands to MySQL but ignores the failure status
returned by the API class MySqlUtil. This made tracking down the error very
difficult.
How-To-RepeatHere is a suggested solution to be added after r2.2:

Provided a local version of the execute function to call the function in the
parent class and check the response and report it to System.error.

Replace several calls to super.execute with this.execute to get this checking
behavior for all accesses to this function. The reasons for the use of super
in this situation was unclear.

Here is the required code:
// ================================================|Public Method Header|====
/**
* Method execute
* An overide of the MySqlUtil method that checks for success and reports
* failure
*
* @param statement A SQL statement for the database to execute.
* @return success The value return from the super class.
*/
// ================================================|Public Method Header|====
public boolean execute( String statement )
{
boolean success = false;

if ( success = super.execute(statement)) {
return success;
}
else {
System.out.println("Execution of the following command "+
"failed: " + statement);
return success;
}
}

Notes---
ID46
CategorymmtxT
SynopsisPermission Problem--MMTx hangs if not root or mmtx user
Severityserious
Priorityhigh
Responsibleguy
Stateclosed
Classsw-bug
Arrival-DateFri Oct 04 22:07:00 EDT 2002
Closed-DateTue Nov 23 11:14:51 EST 2004
Release2.2.A
EnvironmentOperating System: MMTx: Linux MySQL:Linux
OS Version: RH 7.3
Java Version: Java 1.3
MySQL Version:
DescriptionUpdated to 2.2.A, deleted old Tables during update.
When running mmtx from command line, processing hangs silently after output line "Processing 'user-enter phrase'" . . . if the command is run by a user other than mmtx or root [software installed under mmtx user account].

No error messages appear using -w or -9
Had initially installed previous version with mmtxUser to have password mmtx.
Looking at the config.log file for 2.2.a, no password appears for the mmtxUser.
Upgrade to 2.2.a was uneventful--tests at end of install were successful.
How-To-RepeatRun mmtx from an account other than root or the mmtx account where the software is installed.
NotesMMTx 2.2.C contains fix to JDBM which now ignores the
transaction log.
In Version 2.4, MySQL has been replaced with BerkeleyBtrees
ID47
CategorymmtxT
SynopsismmtxUser password clash when running both 2.0 and 2.2
Severitynon-critical
Prioritylow
Responsiblejim
Stateclosed
Classsw-bug
Arrival-DateTue Oct 08 14:50:01 EDT 2002
Closed-DateTue Dec 03 11:03:16 EST 2002
Release2.0, 2.2
EnvironmentOperating System: MMTx: Linux MySQL:Linux
OS Version: RH 7.3
Java Version: Java 1.3
MySQL Version:
Description2.0 has no password policy, 2.2 requires mmtxUser of 8 chars.
2.0 will not work after a 2.2 install if a user changes from having a password to no password, or has a password less than 8 chars and wants to maintain a password. Fix is to update the password in the 2.0 mmtxRegistry.
How-To-RepeatInstall 2.2, change password, or select no password when password exists for 2.0. 2.0 will be broken.
Notes---
ID48
CategorymmtxT
SynopsisInstall script can not connect to the databases.
Severitycritical
Priorityhigh
Responsiblewill
Stateclosed
Classsw-bug
Arrival-DateFri Oct 18 15:41:00 EDT 2002
Closed-DateTue Jan 07 09:39:41 EST 2003
Release2.2.A
EnvironmentOperating System: MMTx: Linux MySQL:Linux
OS Version: Linux 2.4.19
Java Version: Java 1.3
MySQL Version: 3.23.52
DescriptionI can not install metamap! install script fails at the point given below. (my mysqld deamon is working and I am using the correct root password)

-----------

Creating & loading necessary databases ...

java.lang.Exception: Not able to connect to the database.
Check the following common causes:
1) Is the database server (deamon) running?
2) Does the root have a password, and if so,
was it correctly specified?

at gov.nih.nlm.nls.mmtx.install.MMTxInstall.createDatabases(MMTxInstall.java:812)
at gov.nih.nlm.nls.mmtx.install.MMTxInstall.main(MMTxInstall.java:1736)

Installation is not complete: Not able to connect to the database.
Check the following common causes:
1) Is the database server (deamon) running?
2) Does the root have a password, and if so,
was it correctly specified?
How-To-Repeat---
Notes---
ID49
CategorymmtxT
SynopsisMMTx Trouble Report
Severityserious
Prioritymedium
Responsiblejim
Stateclosed
Classsw-bug
Arrival-DateSun Oct 20 20:09:00 EDT 2002
Closed-DateMon Oct 21 07:39:51 EDT 2002
Release---
EnvironmentOperating System: MMTx: MySQL:
OS Version:
Java Version:
MySQL Version:
Description---
How-To-Repeat---
Notes---
ID50
CategorymmtxT
Synopsisunable to properly install MMTx on Windows Me
Severitycritical
Priorityhigh
Responsiblecliff
Stateclosed
Classsw-bug
Arrival-DateWed Oct 30 20:09:00 EST 2002
Closed-DateFri Nov 08 11:15:01 EST 2002
Releasenot running... downloaded mmtx_V2[1].A.jar
EnvironmentOperating System: MMTx: Windows NT/2000 MySQL:Win