xTc @ s d Z d d l m Z d d l Z d d l m Z d d l Z d Z e d Z e
d d e e d d e d Z
d Z d Z e j d
e
d Z d Z d
Z d S( s
Module for querying the RCSB web server using the REST API, as described here:
http://www.rcsb.org/pdb/software/rest.do
There is some overlap with iotbx.pdb.fetch, which really should have gone here
instead, but this module is intended to be used in higher-level automation
pipelines.
i( t divisionN( t parseStrings http://www.rcsb.org/pdb/rests /searchc C s | d k r t j j } n g } | d k r@ | j | n | d IJ| r | d IJd d }
| rw |
d 7}
n | j |
n | d k s | d k r=d } | d d 7} | d k r | d k s t | d
| IJ| d | 7} n | d k r-| d k st | d | IJ| d
| 7} n | j | n | ra| d IJ| j d d n | d k r| d t | IJ| d k r| d k st | j d d t | n t | d k rt d n d } d }
x, | D]$ } | d |
| f 7} |
d 7}
qWd | } t | } t
} | rR| d 7} | d IJn | d IJt j j | | j } | j
S( s
Generate the full XML for a multi-part query with generic search options,
starting from the basic query passed by another function, post it to the
RCSB's web service, and return a list of matching PDB IDs.
Parameters
----------
query_xml : str or None
xray_only : bool, optional
d_max : float, optional
d_min : float, optional
protein_only : bool, optional
data_only : bool, optional
identity_cutoff : int, optional
Apply sequence filtering to remove structures with greater than this
percentage sequence identity. Must be one of 100, 95, 90, 70, 50, 40, 30.
log : file, optional
sort_by_resolution : bool, optional
s Setting up RCSB server query:s limiting to X-ray structuress9 org.pdb.query.simple.ExpTypeQuery
s@ X-RAYsQ
Ys; org.pdb.query.simple.ResolutionQuerys)
betweens" i s' applying resolution cutoff d_min > %fs8
%fs' applying resolution cutoff d_min < %fs8
%fs excluding non-protein modelss; org.pdb.query.simple.ChainTypeQuery
s$ Ys. filtering based on sequence identity < %d %%id sD org.pdb.query.simple.HomologueReductionQuery
s# %ds No queries specified!t s
%d
and
%s
i sB
%s
s /?sortfield=Resolutions will sort by resolutions executing HTTP request...N( t Nonet libtbxt utilst null_outt appendt AssertionErrort intt lent RuntimeErrorR t
url_searcht urlopent readt
splitlines( t query_xmlt xray_onlyt d_maxt d_mint protein_onlyt data_onlyt identity_cutofft logt sort_by_resolutiont queriest
xray_queryt base_clauset queries_stringt kt clauset query_strt parsedt url_finalt result( ( sK /public2/ftp/pub/ccp4/6.5/ccp4-6.5/lib/py2/mmtbx/wwpdb/rcsb_web_services.pyt
post_query sr
c K s | j d d } | j d d } | j d d } | d k rO | d 9} n | d k sa t d | | | | | | f } t | | S(
Nt search_typet blastt expectg{Gz?t min_identityg i id t fastat psiblasts org.pdb.query.simple.SequenceQuery
Sequence Search (expect = %g, search = %s)
%s
%g
%g
%s( R% R( R) ( t popR R# ( t sequencet kwdsR$ R& R' R ( ( sK /public2/ftp/pub/ccp4/6.5/ccp4-6.5/lib/py2/mmtbx/wwpdb/rcsb_web_services.pyt sequence_searchm s
c K sv d t | k o d k n s( t | j d d } | d k sL t d | } d | | | f } t | | S(
sH
Find all entry IDs with the specified chemical ID.
Parameters
----------
resname : str
kwds : dict
Keyword arguments passed to post_query.
Examples
--------
>>> from mmtbx.wwpdb.rcsb_web_services import chemical_id_search
>>> len(chemical_id_search("ZN", data_only=True, identity_cutoff=70))
2874
i i t polymeric_typet Anyt Freet Polymerics! %ss org.pdb.query.simple.ChemCompIdQuery
Chemical ID: %s
%s
%s( R/ R0 R1 ( R
R R* R# ( t resnameR, R. t
polymer_limitR ( ( sK /public2/ftp/pub/ccp4/6.5/ccp4-6.5/lib/py2/mmtbx/wwpdb/rcsb_web_services.pyt chemical_id_search s (
t dimStructurec C s t | d k s t t | d k r. g Sd } | d d j | } d g | } | d d j | 7} t j j | j } t | } g }
t g } | j d } x | D] }
|
j d | } t | d k s t |
j
| d j d j } | j
| | g } xZ | D]R } |
j d
| | f } t | d k s`t | j | d j d j q)W|
j | q W| rt | | } t | d k r| d t | IJ| d d
j t t | IJqn |
S( sy Given a list of PDB IDs and a list of attribute identifiers, returns a
Python list of lists for the IDs and attributes.i s* http://www.rcsb.org/pdb/rest/customReport?s pdbids=%st ,t structureIds &customReportColumns=%st records %s.structureIdi s %s.%ss( WARNING: missing report info for %d IDs:s %st ( R
R t joinR R R
R R t sett getElementsByTagNamet toxmlt
childNodest datat addR t sortedt list( t pdb_idst columnsR t prefixt check_for_missingt url_baset urlt all_columnsR" t xmlrect tablet
report_idst recordsR8 t pdb_id_nodest pdb_idt rowt col_namet row_colt missing_ids( ( sK /public2/ftp/pub/ccp4/6.5/ccp4-6.5/lib/py2/mmtbx/wwpdb/rcsb_web_services.pyt get_custom_report_table s:
$
)c C s t | d d g S( NRD t highResolutionLimit( RT ( RC ( ( sK /public2/ftp/pub/ccp4/6.5/ccp4-6.5/lib/py2/mmtbx/wwpdb/rcsb_web_services.pyt" get_high_resolution_for_structures s c C s. t | d d d d d d d g d d d
t S( sW Return a list of ligands in the specified structure(s), including the
SMILES strings.RD t chainIdt ligandIdt ligandMolecularWeightt
ligandFormulat
ligandNamet ligandSmilesRE t dimEntityRF ( RT t False( RC ( ( sK /public2/ftp/pub/ccp4/6.5/ccp4-6.5/lib/py2/mmtbx/wwpdb/rcsb_web_services.pyt get_ligand_info_for_structures s
( t __doc__t
__future__R t libtbx.utilsR t xml.dom.minidomR t sysRG R t TrueR R^ R# R- R4 t stdoutRT RV R_ ( ( ( sK /public2/ftp/pub/ccp4/6.5/ccp4-6.5/lib/py2/mmtbx/wwpdb/rcsb_web_services.pyt s
X .