medic.1

.\" Man page generated from reStructuredText.
.
.TH MEDIC  "" "" ""
.SH NAME
medic \- a command-line tool to manage a mirror of MEDLINE
.
.nr rst2man-indent-level 0
.
.de1 rstReportMargin
\\$1 \\n[an-margin]
level \\n[rst2man-indent-level]
level margin: \\n[rst2man-indent\\n[rst2man-indent-level]]
-
\\n[rst2man-indent0]
\\n[rst2man-indent1]
\\n[rst2man-indent2]
..
.de1 INDENT
.\" .rstReportMargin pre:
. RS \\$1
. nr rst2man-indent\\n[rst2man-indent-level] \\n[an-margin]
. nr rst2man-indent-level +1
.\" .rstReportMargin post:
..
.de UNINDENT
. RE
.\" indent \\n[an-margin]
.\" old: \\n[rst2man-indent\\n[rst2man-indent-level]]
.nr rst2man-indent-level -1
.\" new: \\n[rst2man-indent\\n[rst2man-indent-level]]
.in \\n[rst2man-indent\\n[rst2man-indent-level]]u
..
.sp
The Swiss Army knife to parse \fI\%MEDLINE\fP XML files or
download eUtils\(aq \fI\%PubMed\fP XML records,
bootstrapping a local MEDLINE/PubMed database,
updating and/or deleting the records, and
writing the contents of selected PMIDs into flat\-files.
A more in\-depth tutorial for this tool is part of my \fI\%blog\fP\&.
\fI\%[image: DOI:10.5281/zendo.9968/https://zenodo.org/badge/4016/fnl/medic.png]
\fP
.sp
DOI:10.5281/zendo.9968
.SH SYNOPSIS
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
medic [options] CMD FILE|PMID...

man medic
medic \-\-help
medic \-\-output /tmp parse baseline/medline*.xml.gz
medic \-\-update parse update/medline*.xml.gz
medic \-\-info delete delete.txt
medic \-\-url sqlite:///tmp.db insert pubmed.xml
medic \-\-pmid\-lists update pmids_to_fetch_online.txt
medic \-\-all update medline13n1000.xml
medic \-\-format html write 1028734 1298474 > out.html
medic \-\-logfile log.txt write pmid_list.txt
.ft P
.fi
.UNINDENT
.UNINDENT
.SH SETUP
.sp
If you are \fBnot\fP using \fBpip install medic\fP, install all
dependencies/requirements:
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
pip install sqlalchemy
# only if using python3 < 3.2:
pip install argparse
.ft P
.fi
.UNINDENT
.UNINDENT
.sp
Install the \fBDB driver\fP you prefer to use (supported are PostgreSQL
and SQLite, with the latter part of the Python StdLib):
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
pip install psycopg2
.ft P
.fi
.UNINDENT
.UNINDENT
.sp
Create the PostgreSQL database:
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
createdb medline
.ft P
.fi
.UNINDENT
.UNINDENT
.sp
If you are fine working with SQLite, you only need to use the path to the
SQLite DB file in the URL option (that will implicitly "create" the DB):
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
medic insert \-\-url sqlite:///tmp.db 123456
.ft P
.fi
.UNINDENT
.UNINDENT
.sp
To \fIcreate the tables in an empty DB\fP, you can just "try" to fetch a record:
As the DB is empty, this will not write anything, but SQL Alchemy will create
the tables for you in the DB:
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
medic write 123 # for PostgreSQL
medic \-\-url sqlite:///tmp.db write 123 # for SQLite
.ft P
.fi
.UNINDENT
.UNINDENT
.SH DESCRIPTION
.sp
\fBmedic [options] COMMAND PMID|FILE...\fP
.sp
The \fB\-\-url URL\fP option represents the DSN of the database and might
be needed (default: \fBpostgresql://localhost/medline\fP); For example:
.INDENT 0.0
.TP
.B PostgreSQL
\fBpostgresql://host//dbname\fP
.TP
.B SQLite DB
\fBsqlite:////absolute/path/to/foo.db\fP or
\fBsqlite:///relative/path/to/foo.db\fP
.UNINDENT
.sp
The five \fBCOMMAND\fP arguments:
.INDENT 0.0
.TP
.B \fBinsert\fP
Create records in the DB by parsing MEDLINE XML files or
by downloading PubMed XML from NCBI eUtils for a list of PMIDs.
The insert fails if the record already exists in the DB.
.TP
.B \fBwrite\fP *
Write records as \fI\%MEDLINE\fP files to a directory, each file named as
"<pmid>.txt". Alternatively, just the TIAB (title and abstract) plain\-text
can be output, and finally, a single file in TSV or HTML format can be
generated (see option \fB\-\-format\fP).
If the requested PMID does not exist in the DB, the command does not fail,
but the relevant file, row, or element will not have been written.
.TP
.B \fBupdate\fP
Insert or update records in the DB (instead of creating them); note that
if a record exists, but is added with \fBcreate\fP, this would throw an
\fIIntegrityError\fP\&. If you are not sure if the records are in the DB or
not, use \fBupdate\fP (N.B. that \fBupdate\fP is slower).
.TP
.B \fBdelete\fP *
Delete records from the DB for a list of PMIDs (using \fB\-\-pmid\-lists\fP)
.TP
.B \fBparse\fP
Does not interact with the DB, but rather creates ".tab" files for each
table that later can be used to load a database, particularly useful when
bootstrapping a large collection.
.UNINDENT
.sp
* Note that \fBwrite\fP and \fBdelete\fP can only use PMID lists (option
\fB\-\-pmid\-lists\fP), so for these two commands, that option is always active
(implicitly).
.sp
For example, to download two PubMed records by PMID and update them in
the DB:
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
medic update 100000 123456
.ft P
.fi
.UNINDENT
.UNINDENT
.sp
Add a single MEDLINE or PubMed XML file to the database:
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
medic insert pudmed.xml
.ft P
.fi
.UNINDENT
.UNINDENT
.sp
Note that if the suffix ".gz" is present, the parser automatically
decompresses the XML file(s) first. This feature \fIonly\fP works with
GNU\-zipped files and the ".gz" suffix must be present.
.sp
Export a few records from the database as HTML (to "articles.html"):
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
medic write \-\-format html \-\-output articles.html 292837491 128374 213487
.ft P
.fi
.UNINDENT
.UNINDENT
.sp
Write the content of the entire database (in MEDLINE format, to STDOUT):
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
medic write ALL
.ft P
.fi
.UNINDENT
.UNINDENT
.sp
Therefore, command line arguments are treated as follows:
.INDENT 0.0
.TP
.B integer values
are always treated as PMIDs to download PubMed XML data
.TP
.B all other values
are always treated as MEDLINE XML files to parse
\fBunless\fP you use the option \fB\-\-pmid\-lists\fP
.TP
.B files ending in ".gz"
are treated as gzipped MEDLINE XML files
.UNINDENT
.SH REQUIREMENTS
.INDENT 0.0
.IP \(bu 2
Python 3.2+
.IP \(bu 2
SQL Alchemy 0.8+
.IP \(bu 2
PostgreSQL 8.4+ or SQLite 3.7+
.UNINDENT
.sp
\fINote\fP that while any DB supported by SQL Alchemy should work, all other DBs
are \fBuntested\fP\&.
.SH LOADING MEDLINE
.sp
Please be aware that the MEDLINE distribution \fBis not unique\fP, meaning that
it contains a few records multiple times (see the section about
\fBVersion IDs\fP).
.sp
Parsing and loading the baseline into a PostgreSQL DB on the same machine:
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
medic parse baseline/medline14n*.xml.gz

for table in citations abstracts authors chemicals databases \e
descriptors identifiers keywords publication_types qualifiers sections;
  do psql medline \-c "COPY $table FROM \(aq\(gapwd\(ga/${table}.tab\(aq;";
done
.ft P
.fi
.UNINDENT
.UNINDENT
.sp
For the update files, you need to go \fIone\-by\-one\fP, adding each one \fIin order\fP,
and using the flag \fB\-\-update\fP when parsing the XML. After parsing an XML file
and \fIbefore\fP loading the dump, run \fBmedic delete delete.txt\fP to get rid of
all entities that will be updated or should be removed (PMIDs listed as
\fBDeleteCitation\fPs):
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
# parse a MEDLINE update file:
medic \-\-update parse medline14n1234.xml.gz

# delete its updated and DeleteCitation records:
medic delete delete.txt

# load (COPY) all tables for that MEDLINE file:
for table in citations abstracts authors chemicals databases \e
descriptors identifiers keywords publication_types qualifiers sections;
  do psql medline \-c "COPY $table FROM \(aq\(gapwd\(ga/${table}.tab\(aq;";
done
.ft P
.fi
.UNINDENT
.UNINDENT
.sp
Alternatively \- simpler but slower \- you can just \fBupdate\fP from the XML
directly:
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
medic update medline14n1234.xml.gz
.ft P
.fi
.UNINDENT
.UNINDENT
.SH VERSION IDS
.sp
MEDLINE has began to use versions to allow publishers to add multiple citations
for the same PMID. This only occurs with 71 articles from one journal,
"PLOS Curr", in the 2013 baseline, creating a total of 149 non\-unique records.
.sp
As this is the only journal and as there may only be one record per PMID in the
database, alternative versions are currently being ignored. In other words, if
a MedlineCitation has a VersionID value other than "1", those records can be
skipped to avoid DB errors from non\-unique records.
.sp
For example, in the 2013 baseline, PMID 20029614 is present ten times in the
baseline, each version at a different stage of revision. Because it is the
first entry (in the order they appear in the baseline files) without a
\fBVersionID\fP or a version of "1" that is the relevant record, \fBmedic\fP by
default filters citations with other versions than "1". If you do want to
process other versions of a citation, use the option \fB\-\-all\fP\&.
.sp
To summarize, \fImedic\fP by default \fBremoves\fP alternate citations.
.SH DATABASE TABLES
.INDENT 0.0
.TP
.B Citation (citations)
\fBpmid\fP:BIGINT, \fIstatus\fP:ENUM(state), \fIyear\fP:SMALLINT,
\fItitle\fP:TEXT, \fIjournal\fP:VARCHAR(256),
\fIpub_date\fP:VARCHAR(256), issue:VARCHAR(256), pagination:VARCHAR(256),
\fIcreated\fP:DATE, completed:DATE, revised:DATE, modified:DATE
.TP
.B Abstract (abstracts)
\fBpmid\fP:FK(Citation), \fBsource\fP:ENUM(type), copyright:TEXT
.TP
.B Section (sections)
\fBpmid\fP:FK(Citation), \fBsource\fP:ENUM(type), \fBseq\fP:SMALLINT,
\fIname\fP:ENUM(section), label:VARCHAR(256), \fIcontent\fP:TEXT, \fItruncated\fP:BOOLEAN
.TP
.B Author (authors)
\fBpmid\fP:FK(Citation), \fBpos\fP:SMALLINT, \fIname\fP:TEXT,
initials:VARCHAR(128), forename:VARCHAR(128), suffix:VARCHAR(128),
.TP
.B PublicationType (publication_types)
\fBpmid\fP:FK(Citation), \fBvalue\fP:VARCHAR(256)
.TP
.B Descriptor (descriptors)
\fBpmid\fP:FK(Citation), \fBnum\fP:SMALLINT, major:BOOL, \fIname\fP:TEXT
.TP
.B Qualifier (qualifiers)
\fBpmid\fP:FK(Descriptor), \fBnum\fP:FK(Descriptor), \fBsub\fP:SMALLINT,
major:BOOL, \fIname\fP:TEXT
.TP
.B Identifier (identifiers)
\fBpmid\fP:FK(Citation), \fBnamespace\fP:VARCHAR(32), \fIvalue\fP:VARCHAR(256)
.TP
.B Database (databases)
\fBpmid\fP:FK(Citation), \fBname\fP:VARCHAR(32), \fBaccession\fP:VARCHAR(256)
.TP
.B Chemical (chemicals)
\fBpmid\fP:FK(Citation), \fBidx\fP:VARCHAR(32), uid:VARCHAR(256),
\fIname\fP:VARCHAR(256)
.TP
.B Keyword (keywords)
\fBpmid\fP:FK(Citation), \fBowner\fP:ENUM(owner), \fBcnt\fP:SMALLINT,
major:BOOL, \fIvalue\fP:TEXT
.UNINDENT
.INDENT 0.0
.IP \(bu 2
\fBbold\fP (Composite) Primary Key
.IP \(bu 2
\fIitalic\fP NOT NULL (Strings that may not be NULL are also never empty.)
.UNINDENT
.SH SUPPORTED XML ELEMENTS
.SS Entities
.INDENT 0.0
.IP \(bu 2
MedlineCitation and ArticleTitle (\fBCitation\fP and \fBIdentifier\fP)
.IP \(bu 2
Abstract and OtherAbstract (\fBAbstract\fP and \fBSection\fP)
.IP \(bu 2
Author (\fBAuthor\fP)
.IP \(bu 2
Chemical (\fBChemical\fP)
.IP \(bu 2
DataBank (\fBDatabase\fP)
.IP \(bu 2
Keyword (\fBKeyword\fP)
.IP \(bu 2
MeshHeading (\fBDescriptor\fP and \fBQualifier\fP)
.IP \(bu 2
PublicationType (\fBPublicationType\fP)
.IP \(bu 2
DeleteCitation (for deleting records when parsing updates)
.UNINDENT
.SS Fields/Values
.INDENT 0.0
.IP \(bu 2
Abstract (with "NLM" as \fBAbstract.source\fP)
.IP \(bu 2
AbstractText (\fBSection.name\fP "Abstract" or the \fINlmCategory\fP, \fBSection.content\fP with \fILabel\fP as \fBSection.label\fP)
.IP \(bu 2
AccessionNumber (\fBDatabase.accession\fP)
.IP \(bu 2
ArticleId (\fBIdentifier.value\fP with \fIIdType\fP as \fBIdentifier.namesapce\fP; only available in online PubMed XML)
.IP \(bu 2
ArticleTitle (\fBCitation.title\fP; if empty, use the VernacularTitle or set to "UNKNOWN")
.IP \(bu 2
CollectiveName (\fBAuthor.name\fP)
.IP \(bu 2
CopyrightInformation (\fBAbstract.copyright\fP)
.IP \(bu 2
DataBankName (\fBDatabase.name\fP)
.IP \(bu 2
DateCompleted (\fBCitation.completed\fP)
.IP \(bu 2
DateCreated (\fBCitation.created\fP)
.IP \(bu 2
DateRevised (\fBCitation.revised\fP)
.IP \(bu 2
DescriptorName (\fBDescriptor.name\fP with \fIMajorTopicYN\fP as \fBDescriptor.major\fP)
.IP \(bu 2
ELocationID (\fBIdentifier.value\fP with \fIEIdType\fP as \fBIdentifier.namespace\fP)
.IP \(bu 2
ForeName (\fBAuthor.forename\fP)
.IP \(bu 2
Initials (\fBAuthor.initials\fP)
.IP \(bu 2
Issue (\fBCitation.issue\fP)
.IP \(bu 2
Keyword (\fBKeyword.value\fP with \fIOwner\fP as \fBKeyword.owner\fP and \fIMajorTopicYN\fP as \fBKeyword.major\fP)
.IP \(bu 2
LastName (\fBAuthor.name\fP)
.IP \(bu 2
MedlineCitation (with \fIStatus\fP as \fBCitation.status\fP)
.IP \(bu 2
MedlineTA (\fBCitation.journal\fP)
.IP \(bu 2
NameOfSubstance (\fBChemical.name\fP)
.IP \(bu 2
MedlinePgn (\fBCitation.pagination\fP)
.IP \(bu 2
OtherAbstract (with \fIType\fP as \fBAbstract.source\fP)
.IP \(bu 2
OtherID (\fBIdentifier.value\fP iff \fISource\fP is "PMC" with \fBIdentifier.namespace\fP as "pmc")
.IP \(bu 2
PMID (\fBCitation.pmid\fP)
.IP \(bu 2
PubDate (\fBCitation.pub_date\fP and \fBCitation.year\fP)
.IP \(bu 2
PublicationType (\fBPublicationType.value\fP)
.IP \(bu 2
QualifierName (\fBQualifier.name\fP with \fIMajorTopicYN\fP as \fBQualifier.major\fP)
.IP \(bu 2
RegistryNumber (\fBChemical.uid\fP)
.IP \(bu 2
Suffix (\fBAuthor.suffix\fP)
.IP \(bu 2
VernacularTitle (\fBCitation.title\fP if AbstractTitle is empty)
.IP \(bu 2
Volume (\fBCitation.issue\fP)
.UNINDENT
.SH VERSION HISTORY
.INDENT 0.0
.TP
.B 2.4.1
.INDENT 7.0
.IP \(bu 2
Fixed a bug while parsing "PubDate" records that asserted the wrong length.
Thanks to Josef Seiser for reporting it.
.UNINDENT
.TP
.B 2.4.0
.INDENT 7.0
.IP \(bu 2
Added a \fByear\fP column to table \fBcitations\fP to make it easier to select for
the year of publication. To migrate/update your DB, run these two commands:
.INDENT 2.0
.INDENT 3.5
.sp
.nf
.ft C
ALTER TABLE citations ADD "year" smallint NOT NULL DEFAULT 0;
UPDATE citations SET year = substring(pub_date FROM \(aq^[12][890][0\-9][0\-9]\(aq)::int;
.ft P
.fi
.UNINDENT
.UNINDENT
.sp
Note that the (currently) oldest publications in MEDLINE are from 1809,
from the first issue of the journal "Med Chir Trans".
.UNINDENT
.TP
.B 2.3.1
.INDENT 7.0
.IP \(bu 2
Fixed a bug when fetching all records on PubMed from database with "ALL".
.UNINDENT
.TP
.B 2.3.0
.INDENT 7.0
.IP \(bu 2
Added special argument "ALL" to write/delete all records.
.UNINDENT
.TP
.B 2.2.0
.INDENT 7.0
.IP \(bu 2
Ensured compatibility with the MEDLINE DTD updates for 2015 and cleaned some
code pieces.
Includes a fix for the bad (long) keyword in PMID 25114415 (with a carriage
return).
.IP \(bu 2
A column was added to the ORM, resulting in backwards incompatible change:
From this version on, the trailing string "\fB(ABSTRACT TRUNCATED AT xxx
WORDS)\fP" is stripped from AbstractText and instead the flag \fBtruncated\fP
has be added to table \fBsections\fP and is set if the string was present (but
has been removed); To migrate your Postgres database, please run:
.INDENT 2.0
.INDENT 3.5
.sp
.nf
.ft C
ALTER TABLE sections ADD "truncated" boolean NOT NULL DEFAULT \(aqfalse\(aq;
.ft P
.fi
.UNINDENT
.UNINDENT
.sp
Note that for one case this produces a Section with just one whitespace
character, because the original content was only the "ABSTRACT TRUNCATED..."
message.
.IP \(bu 2
MEDLINE formatted output is now written to STDOUT or a single file, because
it makes selecting specific fields with grep very easy. Records are
separated with an empty line.
.UNINDENT
.TP
.B 2.1.7
.INDENT 7.0
.IP \(bu 2
Work\-around for the limit of SQLite that only lets you use 999 variables per
query.
.IP \(bu 2
Corrected the outdated VernacularTitle documentation in this document.
.UNINDENT
.TP
.B 2.1.6
.INDENT 7.0
.IP \(bu 2
Work\-around for parsing citations that have an empty ArticleTitle element
(which they shouldn\(aqt, according to the DTD): Either use the VernacularTitle
(e.g., PMID 22536004), or otherwise set the title to "UNKNOWN" (the empty
string is not a valid title) and log a warning.
.IP \(bu 2
Work\-around for non\-unique PublicationType entries (e.g., PMID 10500000):
drop non\-unique PublicationTypes (with the same PMID and value).
.IP \(bu 2
Corrected left\-over "Medline" entity names in this document to "Citation".
.UNINDENT
.TP
.B 2.1.5
.INDENT 7.0
.IP \(bu 2
Added page_size=MAX and synchronous=OFF pragmas for SQLite DBs (hat\-tip to
Jason)
.UNINDENT
.TP
.B 2.1.4
.INDENT 7.0
.IP \(bu 2
A MEDLINE issue found by Jason: PMID 24073073 has an empty keyword and
keyword list that should not be there; Medic prevents adding improper data
to the DB by raising an AssertionError.
To deal with such cases, medic now ensures each keyword is non\-empty before
attempting to generate a database entry and drops empty (i.e.,
whitespace\-only) keyword data.
.UNINDENT
.TP
.B 2.1.3
.INDENT 7.0
.IP \(bu 2
Jason Hennessey changed the (absolute) path of medic\(aqs man\-page, set to
\fB/usr/local/share/man\fP in the setup script, to a relative location
(\fBshare/man\fP) to avoid issues when installing medic in a virtualenv
.UNINDENT
.TP
.B 2.1.2
.INDENT 7.0
.IP \(bu 2
fixed a bug where SQLite did not find the implicit FK\->PK reference
(thanks to Jason Hennessey for reporting the issue)
.UNINDENT
.TP
.B 2.1.1
.INDENT 7.0
.IP \(bu 2
added SQLite temporary DB example URL to help output
.IP \(bu 2
refactored HTML output code
.UNINDENT
.TP
.B 2.1.0
.INDENT 7.0
.IP \(bu 2
DB schema change from: \fBrecords() \-> sections(content)\fP
to: \fBcitations(title) \-> abstracts(copyright) \-> sections(content)\fP
.IP \(bu 2
name change: the entity/table Medline/records is now called
Citation/citations
.IP \(bu 2
title and copyright text is no longer stored in Section/sections
.IP \(bu 2
added a new Abstract/abstracts entity/table with a \fBcopyright\fP attribute
(formerly stored in \fBsections.content\fP with \fBname\fP = \(aqCopyright\(aq)
.IP \(bu 2
added a new \fBcitations.title\fP attribute
(formerly stored in \fBsections.content\fP with \fBname\fP = \(aqTitle\(aq)
.IP \(bu 2
added a new \fBsource\fP primary\-key attribute to Section and Abstract
(set to either \(aqNLM\(aq for regular Abstract elements or to
the value of the OtherAbstract Type attribute for other abstracts)
.IP \(bu 2
skipping "Abstract available from the publisher."\-only abstracts
.UNINDENT
.TP
.B 2.0.2
.INDENT 7.0
.IP \(bu 2
made the use of \fB\-\-pmid\-lists\fP for \fBdelete\fP and \fBwrite\fP implicit
.IP \(bu 2
added instructions to bootstrap the tables in a PostgreSQL DB
.IP \(bu 2
minor improvements to this manual
.IP \(bu 2
fixed a bug when inserting/updating from MEDLINE XML files
.UNINDENT
.TP
.B 2.0.1
.INDENT 7.0
.IP \(bu 2
fixed a bug that lead to skipping of abstracts
(thanks to Chris Roeder for detecting the issue)
.UNINDENT
.TP
.B 2.0.0
.INDENT 7.0
.IP \(bu 2
added Keywords and PublicationTypes
.IP \(bu 2
added MEDLINE publication date, volume, issue, and pagination support
.IP \(bu 2
added MEDLINE output format and made it the default
.IP \(bu 2
DB structure change: descriptors.major and qualifiers.major columns swapped
.IP \(bu 2
DB structure change: section.name is now an untyped varchar (OtherAbstract
separation)
.IP \(bu 2
cleaned up the ORM test cases
.UNINDENT
.TP
.B 1.1.1
.INDENT 7.0
.IP \(bu 2
code cleanup (PEP8, PyFlake)
.IP \(bu 2
fixed an issue where the parser would not leave the skipping state
.UNINDENT
.TP
.B 1.1.0
.INDENT 7.0
.IP \(bu 2
\fB\-\-update parse\fP now writes a file to use with \fB\-\-pmid\-lists delete\fP
.IP \(bu 2
fixed a bug with CRUD manager
.IP \(bu 2
added a man page
.UNINDENT
.TP
.B 1.0.2
.INDENT 7.0
.IP \(bu 2
fixes to make the PyPi version and \fBpip install medic\fP work
.UNINDENT
.TP
.B 1.0.1
.INDENT 7.0
.IP \(bu 2
updates to the setup.py and README.rst files
.UNINDENT
.TP
.B 1.0.0
.INDENT 7.0
.IP \(bu 2
initial release
.UNINDENT
.UNINDENT
.SH COPYRIGHT AND LICENSE
.sp
License: \fI\%GNU GPL v3\fP\&.
Copyright 2012\-2014 Florian Leitner. All rights reserved.
.\" Generated by docutils manpage writer.
.