From 834a586d3f07f16bff06a20c81199bc4aa3e7957 Mon Sep 17 00:00:00 2001
From: Ira Cooke <iracooke@gmail.com>
Date: Mon, 7 Jul 2014 11:16:09 +1000
Subject: [PATCH] Add sqlite datatype and corresponding dataprovider

---
 datatypes_conf.xml.sample                     |  2 +
 lib/galaxy/datatypes/binary.py                | 43 +++++++++++++++++++
 lib/galaxy/datatypes/dataproviders/dataset.py | 27 ++++++++++++
 3 files changed, 72 insertions(+)

diff --git a/datatypes_conf.xml.sample b/datatypes_conf.xml.sample
index 4ff7d638a21..e2c845278d7 100644
--- a/datatypes_conf.xml.sample
+++ b/datatypes_conf.xml.sample
@@ -177,6 +177,7 @@
     <datatype extension="taxonomy" type="galaxy.datatypes.tabular:Taxonomy" display_in_upload="true"/>
     <datatype extension="tabular" type="galaxy.datatypes.tabular:Tabular" display_in_upload="true" description="Any data in tab delimited format (tabular)." description_url="https://wiki.galaxyproject.org/Learn/Datatypes#Tabular_.28tab_delimited.29"/>
     <datatype extension="twobit" type="galaxy.datatypes.binary:TwoBit" mimetype="application/octet-stream" display_in_upload="true"/>
+    <datatype extension="sqlite" type="galaxy.datatypes.binary:SQlite" mimetype="application/octet-stream" display_in_upload="true"/>    
     <datatype extension="txt" type="galaxy.datatypes.data:Text" display_in_upload="true" description="Any text file." description_url="https://wiki.galaxyproject.org/Learn/Datatypes#Plain_text"/>
     <datatype extension="linecount" type="galaxy.datatypes.data:LineCount" display_in_upload="false"/>
     <datatype extension="memexml" type="galaxy.datatypes.xml:MEMEXml" mimetype="application/xml" display_in_upload="true"/>
@@ -262,6 +263,7 @@
     -->
     <sniffer type="galaxy.datatypes.tabular:Vcf"/>
     <sniffer type="galaxy.datatypes.binary:TwoBit"/>
+    <sniffer type="galaxy.datatypes.binary:SQlite"/>
     <sniffer type="galaxy.datatypes.binary:Bam"/>
     <sniffer type="galaxy.datatypes.binary:Sff"/>
     <sniffer type="galaxy.datatypes.xml:Phyloxml"/>
diff --git a/lib/galaxy/datatypes/binary.py b/lib/galaxy/datatypes/binary.py
index a6aa0e1ec03..b6b7dbca433 100644
--- a/lib/galaxy/datatypes/binary.py
+++ b/lib/galaxy/datatypes/binary.py
@@ -12,6 +12,7 @@ import struct
 import subprocess
 import tempfile
 import zipfile
+import sqlite3
 
 from urllib import urlencode, quote_plus
 from galaxy import eggs
@@ -545,3 +546,45 @@ class TwoBit (Binary):
             return "Binary TwoBit format nucleotide file (%s)" % (data.nice_size(dataset.get_size()))
 
 Binary.register_sniffable_binary_format("twobit", "twobit", TwoBit)
+
+
+@dataproviders.decorators.has_dataproviders
+class SQlite ( Binary ):
+    file_ext = "sqlite"
+
+    # Connects and runs a query that should work on any real database 
+    # If the file is not sqlite, an exception will be thrown and the sniffer will return false
+    def sniff( self, filename ):
+        try:
+            conn = sqlite3.connect(filename)
+            schema_version=conn.cursor().execute("pragma schema_version").fetchone()
+            conn.close()
+            if schema_version is not None:
+                return True
+            return False
+        except:
+            return False
+
+    def set_peek( self, dataset, is_multi_byte=False ):
+        if not dataset.dataset.purged:
+            dataset.peek  = "SQLite Database"
+            dataset.blurb = data.nice_size( dataset.get_size() )
+        else:
+            dataset.peek = 'file does not exist'
+            dataset.blurb = 'file purged from disk'
+
+    def display_peek( self, dataset ):
+        try:
+            return dataset.peek
+        except:
+            return "SQLite Database (%s)" % ( data.nice_size( dataset.get_size() ) )
+
+
+    @dataproviders.decorators.dataprovider_factory( 'sqlite', dataproviders.dataset.SQliteDataProvider.settings )
+    def sqlite_dataprovider( self, dataset, **settings ):
+        dataset_source = dataproviders.dataset.DatasetDataProvider( dataset )
+        return dataproviders.dataset.SQliteDataProvider( dataset_source, **settings )
+
+
+Binary.register_sniffable_binary_format("sqlite","sqlite",SQlite)
+
diff --git a/lib/galaxy/datatypes/dataproviders/dataset.py b/lib/galaxy/datatypes/dataproviders/dataset.py
index 88dab3c102a..cb5c38fd2f3 100644
--- a/lib/galaxy/datatypes/dataproviders/dataset.py
+++ b/lib/galaxy/datatypes/dataproviders/dataset.py
@@ -11,6 +11,7 @@ import base
 import line
 import column
 import external
+import sqlite3
 
 from galaxy import eggs
 eggs.require( 'bx-python' )
@@ -700,3 +701,29 @@ class BGzipTabixDataProvider( base.DataProvider ):
         #TODO: as samtools - need more info on output format
         raise NotImplementedError()
         super( BGzipTabixDataProvider, self ).__init__( dataset, **kwargs )
+
+
+
+class SQliteDataProvider ( base.DataProvider ):
+    """
+    Data provider that uses a sqlite database file as its source.
+
+    Allows any query to be run and returns the resulting rows as sqlite3 row objects
+    """
+    settings = {
+        'query' : 'str'
+    }
+
+    def __init__( self, source, query=None, **kwargs ):
+        self.query=query
+        self.connection =  sqlite3.connect(source.dataset.file_name);
+        self.connection.row_factory = sqlite3.Row
+        super( SQliteDataProvider, self ).__init__( source, **kwargs )
+
+    def __iter__( self ):
+        if self.query is not None:
+            for row in self.connection.cursor().execute(self.query):
+                yield row
+        else:
+            yield
+