Changeset: 837dec34b419 for MonetDB
URL: http://dev.monetdb.org/hg/MonetDB?cmd=changeset;node=837dec34b419
Added Files:
monetdb5/extras/pyapi/Tests/pyapi_modify_input.malC
monetdb5/extras/pyapi/Tests/pyapi_modify_input.stable.err
monetdb5/extras/pyapi/Tests/pyapi_modify_input.stable.out
monetdb5/extras/pyapi/bytearray.c
monetdb5/extras/pyapi/bytearray.h
Modified Files:
monetdb5/extras/pyapi/Makefile.ag
monetdb5/extras/pyapi/Tests/All
monetdb5/extras/pyapi/Tests/pyapi_pandas.stable.err
monetdb5/extras/pyapi/Tests/pyapi_returntypes.stable.err
monetdb5/extras/pyapi/Tests/pyapi_types_huge.stable.err
monetdb5/extras/pyapi/Tests/pyapi_types_numeric.stable.err
monetdb5/extras/pyapi/pyapi.c
monetdb5/extras/pyapi/type_conversion.c
monetdb5/extras/pyapi/type_conversion.h
sql/backends/monet5/Tests/pyapi08.sql
Branch: pyapi
Log Message:
Added the option of converting strings into PyByteArrayObjects instead of
PyStringObjects, which passes a reference to the string rather than copying the
entire string.
diffs (truncated from 663 to 300 lines):
diff --git a/monetdb5/extras/pyapi/Makefile.ag
b/monetdb5/extras/pyapi/Makefile.ag
--- a/monetdb5/extras/pyapi/Makefile.ag
+++ b/monetdb5/extras/pyapi/Makefile.ag
@@ -17,7 +17,7 @@ MTSAFE
lib__pyapi = {
MODULE
DIR = libdir/monetdb5
- SOURCES = pyapi.c pyapi.h unicode.c unicode.h pytypes.c pytypes.h
type_conversion.c type_conversion.h
+ SOURCES = pyapi.c pyapi.h unicode.c unicode.h pytypes.c pytypes.h
type_conversion.c type_conversion.h bytearray.c bytearray.h
XDEPS = $(libpy_LIBDEP)
LIBS = ../../tools/libmonetdb5 \
../../../gdk/libbat \
diff --git a/monetdb5/extras/pyapi/Tests/All b/monetdb5/extras/pyapi/Tests/All
--- a/monetdb5/extras/pyapi/Tests/All
+++ b/monetdb5/extras/pyapi/Tests/All
@@ -7,3 +7,4 @@ HAVE_LIBPY?pyapi_types_string
HAVE_LIBPY?pyapi_numpy_boolean
HAVE_LIBPY?pyapi_types_numeric
HAVE_LIBPY?pyapi_numpy_numeric_nested
+HAVE_LIBPY?pyapi_modify_input
diff --git a/monetdb5/extras/pyapi/Tests/pyapi_modify_input.malC
b/monetdb5/extras/pyapi/Tests/pyapi_modify_input.malC
new file mode 100644
--- /dev/null
+++ b/monetdb5/extras/pyapi/Tests/pyapi_modify_input.malC
@@ -0,0 +1,34 @@
+
+# a number of operations that might not work properly with byte arrays
+# basic strings
+bstr:= bat.new(:oid,:str);
+bat.append(bstr,"asdf":str);
+bat.append(bstr,"sd asd asd asd asd a":str);
+bat.append(bstr,"1234":str);
+bat.append(bstr,"test":str);
+
+# we are not allowed to change the value of the input BAT from within python
code, check if this is possible with byte array operations
+# setitem
+rstr:bat[:oid,:str] :=
pyapi.eval(nil:ptr,"arg1[0].__setitem__(0,'a')\nreturn(arg1)",bstr);
+# subscript
+rstr:bat[:oid,:str] := pyapi.eval(nil:ptr,"arg1[0][0]='d'\nreturn(arg1)",bstr);
+# reverse
+rstr:bat[:oid,:str] :=
pyapi.eval(nil:ptr,"arg1[0].reverse()\nreturn(arg1)",bstr);
+# replace
+rstr:bat[:oid,:str] :=
pyapi.eval(nil:ptr,"arg1[0].replace('a','d')\nreturn(arg1)",bstr);
+# remove
+rstr:bat[:oid,:str] := pyapi.eval(nil:ptr,"a =
arg1[0].remove('a')\nreturn(arg1)",bstr);
+# delete
+rstr:bat[:oid,:str] := pyapi.eval(nil:ptr,"a = arg1[0]\ndel
a\nreturn(arg1)",bstr);
+# pop
+rstr:bat[:oid,:str] := pyapi.eval(nil:ptr,"a =
arg1[0].pop()\nreturn(arg1)",bstr);
+io.print(bstr);
+
+#try the same with other types just to be sure
+a:= bat.new(:oid,:int);
+bat.append(a,1:int);
+bat.append(a,2:int);
+bat.append(a,3:int);
+bat.append(a,4:int);
+result:bat[:oid,:int] := pyapi.eval(nil:ptr, "arg1[0]=4\nreturn(arg1)", a);
+io.print(a);
diff --git a/monetdb5/extras/pyapi/Tests/pyapi_modify_input.stable.err
b/monetdb5/extras/pyapi/Tests/pyapi_modify_input.stable.err
new file mode 100644
--- /dev/null
+++ b/monetdb5/extras/pyapi/Tests/pyapi_modify_input.stable.err
@@ -0,0 +1,58 @@
+stderr of test 'pyapi_modify_input` in directory 'monetdb5/extras/pyapi`
itself:
+
+
+# 12:27:28 >
+# 12:27:28 > "mserver5" "--debug=10" "--set" "gdk_nr_threads=0" "--set"
"mapi_open=true" "--set" "mapi_port=31030" "--set"
"mapi_usock=/var/tmp/mtest-30203/.s.monetdb.31030" "--set" "monet_prompt="
"--forcemito" "--set" "mal_listing=2"
"--dbpath=/home/mytherin/opt/var/mTests_monetdb5_extras_pyapi" "--set"
"mal_listing=2" "--set" "embedded_r=true" "--set" "embedded_py=true"
+# 12:27:28 >
+
+# builtin opt gdk_dbpath = /home/mytherin/opt/var/monetdb5/dbfarm/demo
+# builtin opt gdk_debug = 0
+# builtin opt gdk_vmtrim = no
+# builtin opt monet_prompt = >
+# builtin opt monet_daemon = no
+# builtin opt mapi_port = 50000
+# builtin opt mapi_open = false
+# builtin opt mapi_autosense = false
+# builtin opt sql_optimizer = default_pipe
+# builtin opt sql_debug = 0
+# cmdline opt gdk_nr_threads = 0
+# cmdline opt mapi_open = true
+# cmdline opt mapi_port = 31030
+# cmdline opt mapi_usock = /var/tmp/mtest-30203/.s.monetdb.31030
+# cmdline opt monet_prompt =
+# cmdline opt mal_listing = 2
+# cmdline opt gdk_dbpath = /home/mytherin/opt/var/mTests_monetdb5_extras_pyapi
+# cmdline opt mal_listing = 2
+# cmdline opt embedded_r = true
+# cmdline opt embedded_py = true
+# cmdline opt gdk_debug = 536870922
+
+# 12:27:29 >
+# 12:27:29 > "mclient" "-lmal" "-ftest" "-Eutf-8" "-i" "-e"
"--host=/var/tmp/mtest-30203" "--port=31030"
+# 12:27:29 >
+
+MAPI = (monetdb) /var/tmp/mtest-30203/.s.monetdb.31030
+QUERY = rstr:bat[:oid,:str] :=
pyapi.eval(nil:ptr,"arg1[0][0]='d'\nreturn(arg1)",bstr);
+ERROR = !MALException:pyapi.eval:Python exception
+ !This ByteArray references to a BAT in the database, you may not
assign to it.
+MAPI = (monetdb) /var/tmp/mtest-30203/.s.monetdb.31030
+QUERY = rstr:bat[:oid,:str] :=
pyapi.eval(nil:ptr,"arg1[0].reverse()\nreturn(arg1)",bstr);
+ERROR = !MALException:pyapi.eval:Python exception
+ !This ByteArray references to a BAT in the database, you may not
reverse it.
+MAPI = (monetdb) /var/tmp/mtest-30203/.s.monetdb.31030
+QUERY = rstr:bat[:oid,:str] := pyapi.eval(nil:ptr,"a =
arg1[0].remove('a')\nreturn(arg1)",bstr);
+ERROR = !MALException:pyapi.eval:Python exception
+ !This ByteArray references to a BAT in the database, you may not
remove anything from it.
+MAPI = (monetdb) /var/tmp/mtest-30203/.s.monetdb.31030
+QUERY = rstr:bat[:oid,:str] := pyapi.eval(nil:ptr,"a =
arg1[0].pop()\nreturn(arg1)",bstr);
+ERROR = !MALException:pyapi.eval:Python exception
+ !This ByteArray references to a BAT in the database, you may not
remove anything from it.
+MAPI = (monetdb) /var/tmp/mtest-30203/.s.monetdb.31030
+QUERY = result:bat[:oid,:int] := pyapi.eval(nil:ptr,
"arg1[0]=4\nreturn(arg1)", a);
+ERROR = !MALException:pyapi.eval:Python exception
+ !assignment destination is read-only
+
+# 12:27:29 >
+# 12:27:29 > "Done."
+# 12:27:29 >
+
diff --git a/monetdb5/extras/pyapi/Tests/pyapi_modify_input.stable.out
b/monetdb5/extras/pyapi/Tests/pyapi_modify_input.stable.out
new file mode 100644
--- /dev/null
+++ b/monetdb5/extras/pyapi/Tests/pyapi_modify_input.stable.out
@@ -0,0 +1,54 @@
+stdout of test 'pyapi_modify_input` in directory 'monetdb5/extras/pyapi`
itself:
+
+
+# 12:27:28 >
+# 12:27:28 > "mserver5" "--debug=10" "--set" "gdk_nr_threads=0" "--set"
"mapi_open=true" "--set" "mapi_port=31030" "--set"
"mapi_usock=/var/tmp/mtest-30203/.s.monetdb.31030" "--set" "monet_prompt="
"--forcemito" "--set" "mal_listing=2"
"--dbpath=/home/mytherin/opt/var/mTests_monetdb5_extras_pyapi" "--set"
"mal_listing=2" "--set" "embedded_r=true" "--set" "embedded_py=true"
+# 12:27:28 >
+
+# MonetDB 5 server v11.22.0
+# This is an unreleased version
+# Serving database 'mTests_monetdb5_extras_pyapi', using 8 threads
+# Compiled for x86_64-unknown-linux-gnu/64bit with 64bit OIDs and 128bit
integers dynamically linked
+# Found 7.684 GiB available main-memory.
+# Copyright (c) 1993-July 2008 CWI.
+# Copyright (c) August 2008-2015 MonetDB B.V., all rights reserved
+# Visit http://www.monetdb.org/ for further information
+# Listening for connection requests on mapi:monetdb://mytherin-N750JV:31030/
+# Listening for UNIX domain connection requests on
mapi:monetdb:///var/tmp/mtest-30203/.s.monetdb.31030
+# Start processing logs sql/sql_logs version 52200
+# Start reading the write-ahead log 'sql_logs/sql/log.4'
+# Finished reading the write-ahead log 'sql_logs/sql/log.4'
+# Finished processing logs sql/sql_logs
+# MonetDB/SQL module loaded
+# MonetDB/Python module loaded
+# MonetDB/R module loaded
+
+Ready.
+
+# 12:27:29 >
+# 12:27:29 > "mclient" "-lmal" "-ftest" "-Eutf-8" "-i" "-e"
"--host=/var/tmp/mtest-30203" "--port=31030"
+# 12:27:29 >
+
+#io.print(bstr);
+#--------------------------#
+# h t # name
+# void str # type
+#--------------------------#
+[ 0@0, "asdf" ]
+[ 1@0, "sd asd asd asd asd a" ]
+[ 2@0, "1234" ]
+[ 3@0, "test" ]
+#io.print(a);
+#--------------------------#
+# h t # name
+# void int # type
+#--------------------------#
+[ 0@0, 1 ]
+[ 1@0, 2 ]
+[ 2@0, 3 ]
+[ 3@0, 4 ]
+
+# 12:27:29 >
+# 12:27:29 > "Done."
+# 12:27:29 >
+
diff --git a/monetdb5/extras/pyapi/Tests/pyapi_pandas.stable.err
b/monetdb5/extras/pyapi/Tests/pyapi_pandas.stable.err
--- a/monetdb5/extras/pyapi/Tests/pyapi_pandas.stable.err
+++ b/monetdb5/extras/pyapi/Tests/pyapi_pandas.stable.err
@@ -31,7 +31,7 @@ stderr of test 'pyapi_pandas` in directo
# 19:41:50 > "mclient" "-lmal" "-ftest" "-Eutf-8" "-i" "-e"
"--host=/var/tmp/mtest-7261" "--port=31934"
# 19:41:50 >
-MAPI = (monetdb) /var/tmp/mtest-16516/.s.monetdb.30115
+MAPI = (monetdb) /var/tmp/mtest-30203/.s.monetdb.31030
QUERY = (r:bat[:oid,:int], s:bat[:oid,:int]) := pyapi.eval(nil:ptr,"import
pandas as pd\ndf = pd.DataFrame({\'Group\': arg1, \'Values\': arg2, 'Values2':
arg2})\nreturn(df)", g, c);
ERROR = !MALException:pyapi.eval:An array of size 3 was returned, yet we
expect a list of 2 columns. The result is invalid.
diff --git a/monetdb5/extras/pyapi/Tests/pyapi_returntypes.stable.err
b/monetdb5/extras/pyapi/Tests/pyapi_returntypes.stable.err
--- a/monetdb5/extras/pyapi/Tests/pyapi_returntypes.stable.err
+++ b/monetdb5/extras/pyapi/Tests/pyapi_returntypes.stable.err
@@ -31,34 +31,34 @@ stderr of test 'pyapi_returntypes` in di
# 23:33:07 > "mclient" "-lmal" "-ftest" "-Eutf-8" "-i" "-e"
"--host=/var/tmp/mtest-20340" "--port=36739"
# 23:33:07 >
-MAPI = (monetdb) /var/tmp/mtest-16516/.s.monetdb.30115
+MAPI = (monetdb) /var/tmp/mtest-30203/.s.monetdb.31030
QUERY = r:bat[:oid,:int] := pyapi.eval(nil:ptr,"return(\"Test\")");
ERROR = !MALException:pyapi.eval:Could not convert from type STRING to type int
-MAPI = (monetdb) /var/tmp/mtest-16516/.s.monetdb.30115
+MAPI = (monetdb) /var/tmp/mtest-30203/.s.monetdb.31030
QUERY = (r:bat[:oid,:int], s:bat[:oid,:int]) := pyapi.eval(nil:ptr,"class
NewClass:\n x = 5\n\nreturn(NewClass())");
ERROR = !MALException:pyapi.eval:Unsupported result object. Expected either an
array, a numpy array, a numpy masked array or a pandas data frame, but received
an object of type "<type 'instance'>"
-MAPI = (monetdb) /var/tmp/mtest-16516/.s.monetdb.30115
+MAPI = (monetdb) /var/tmp/mtest-30203/.s.monetdb.31030
QUERY = (r:bat[:oid,:int], s:bat[:oid,:int]) :=
pyapi.eval(nil:ptr,"return(12)");
ERROR = !MALException:pyapi.eval:A single scalar was returned, yet we expect a
list of 2 columns. We can only convert a single scalar into a single column,
thus the result is invalid.
-MAPI = (monetdb) /var/tmp/mtest-16516/.s.monetdb.30115
+MAPI = (monetdb) /var/tmp/mtest-30203/.s.monetdb.31030
QUERY = (r:bat[:oid,:int], s:bat[:oid,:int]) :=
pyapi.eval(nil:ptr,"return(numpy.array([12]))");
ERROR = !MALException:pyapi.eval:A single array was returned, yet we expect a
list of 2 columns. The result is invalid.
-MAPI = (monetdb) /var/tmp/mtest-16516/.s.monetdb.30115
+MAPI = (monetdb) /var/tmp/mtest-30203/.s.monetdb.31030
QUERY = (r:bat[:oid,:int], s:bat[:oid,:int]) :=
pyapi.eval(nil:ptr,"return([12])");
ERROR = !MALException:pyapi.eval:A single array was returned, yet we expect a
list of 2 columns. The result is invalid.
-MAPI = (monetdb) /var/tmp/mtest-16516/.s.monetdb.30115
+MAPI = (monetdb) /var/tmp/mtest-30203/.s.monetdb.31030
QUERY = r:bat[:oid,:int] := pyapi.eval(nil:ptr,"return([[33,24,55],
[44,66,345]])");
ERROR = !MALException:pyapi.eval:An array of size 2 was returned, yet we
expect a list of 1 columns. The result is invalid.
-MAPI = (monetdb) /var/tmp/mtest-16516/.s.monetdb.30115
+MAPI = (monetdb) /var/tmp/mtest-30203/.s.monetdb.31030
QUERY = a:bat[:oid,:str] := pyapi.eval(nil:ptr,"x =
unicode(\"hello\")\nreturn(x.encode(\"utf32\"))");
ERROR = !MALException:pyapi.eval:Invalid string encoding used. Please return a
regular ASCII string, or a Numpy_Unicode object.
-MAPI = (monetdb) /var/tmp/mtest-16516/.s.monetdb.30115
+MAPI = (monetdb) /var/tmp/mtest-30203/.s.monetdb.31030
QUERY = bb:bat[:oid,:int] := pyapi.eval(nil:ptr,"return (1");
ERROR = !MALException:pyapi.eval:Could not parse Python code
! 1. def pyfun():
!> 2. return (1
!invalid syntax (<string>, line 2)
-MAPI = (monetdb) /var/tmp/mtest-16516/.s.monetdb.30115
+MAPI = (monetdb) /var/tmp/mtest-30203/.s.monetdb.31030
QUERY = cc:bat[:oid,:int] := pyapi.eval(nil:ptr,"x = 4\n x++\n\treturn (x)");
ERROR = !MALException:pyapi.eval:Could not parse Python code
! 1. def pyfun():
@@ -66,10 +66,10 @@ ERROR = !MALException:pyapi.eval:Could n
!> 3. x++
! 4. return (x)
!unexpected indent (<string>, line 3)
-MAPI = (monetdb) /var/tmp/mtest-16516/.s.monetdb.30115
+MAPI = (monetdb) /var/tmp/mtest-30203/.s.monetdb.31030
QUERY = (str1:bat[:oid,:str], str2:bat[:oid,:str]) := pyapi.eval(nil:ptr,"x =
unicode(\"hello\")\nreturn(numpy.array([[x.encode(\"utf32\")],
[x.encode(\"utf32\")]]))");
ERROR = !MALException:pyapi.eval:Invalid string encoding used. Please return a
regular ASCII string, or a Numpy_Unicode object.
-MAPI = (monetdb) /var/tmp/mtest-16516/.s.monetdb.30115
+MAPI = (monetdb) /var/tmp/mtest-30203/.s.monetdb.31030
QUERY = (str1:bat[:oid,:str], str2:bat[:oid,:str]) :=
pyapi.eval(nil:ptr,"return(numpy.array([[\"Hëllo\", \"Hello Again\"], [\"Hello
Again Again\",\"That's quite enough.\"]]))");
ERROR = !MALException:pyapi.eval:Invalid string encoding used. Please return a
regular ASCII string, or a Numpy_Unicode object.
diff --git a/monetdb5/extras/pyapi/Tests/pyapi_types_huge.stable.err
b/monetdb5/extras/pyapi/Tests/pyapi_types_huge.stable.err
--- a/monetdb5/extras/pyapi/Tests/pyapi_types_huge.stable.err
+++ b/monetdb5/extras/pyapi/Tests/pyapi_types_huge.stable.err
@@ -26,14 +26,14 @@ stderr of test 'pyapi_types_huge` in dir
# cmdline opt embedded_r = true
# cmdline opt embedded_py = true
# cmdline opt gdk_debug = 536870922
-!WARNING: Type "hge" (128 bit) is unsupported by Numpy. The numbers are
instead converted to python objects of type "long". This is likely very slow.
-!WARNING: Type "hge" (128 bit) is unsupported by Numpy. The numbers are
instead converted to python objects of type "long". This is likely very slow.
-!WARNING: Type "hge" (128 bit) is unsupported by Numpy. The numbers are
instead converted to python objects of type "long". This is likely very slow.
-!WARNING: Type "hge" (128 bit) is unsupported by Numpy. The numbers are
instead converted to python objects of type "long". This is likely very slow.
-!WARNING: Type "hge" (128 bit) is unsupported by Numpy. The numbers are
instead converted to python objects of type "long". This is likely very slow.
-!WARNING: Type "hge" (128 bit) is unsupported by Numpy. The numbers are
instead converted to python objects of type "long". This is likely very slow.
-!WARNING: Type "hge" (128 bit) is unsupported by Numpy. The numbers are
instead converted to python objects of type "long". This is likely very slow.
-!WARNING: Type "hge" (128 bit) is unsupported by Numpy. The numbers are
instead converted to python objects of type "long". This is likely very slow.
+!PERFORMANCE WARNING: Type "hge" (128 bit) is unsupported by Numpy. The
numbers are instead converted to python objects of type "PyLong". This means a
python object is constructed for every huge integer and the entire column is
copied.
+!PERFORMANCE WARNING: Type "hge" (128 bit) is unsupported by Numpy. The
numbers are instead converted to python objects of type "PyLong". This means a
python object is constructed for every huge integer and the entire column is
copied.
+!PERFORMANCE WARNING: Type "hge" (128 bit) is unsupported by Numpy. The
numbers are instead converted to python objects of type "PyLong". This means a
python object is constructed for every huge integer and the entire column is
copied.
+!PERFORMANCE WARNING: Type "hge" (128 bit) is unsupported by Numpy. The
numbers are instead converted to python objects of type "PyLong". This means a
python object is constructed for every huge integer and the entire column is
copied.
+!PERFORMANCE WARNING: Type "hge" (128 bit) is unsupported by Numpy. The
numbers are instead converted to python objects of type "PyLong". This means a
python object is constructed for every huge integer and the entire column is
copied.
+!PERFORMANCE WARNING: Type "hge" (128 bit) is unsupported by Numpy. The
numbers are instead converted to python objects of type "PyLong". This means a
python object is constructed for every huge integer and the entire column is
copied.
+!PERFORMANCE WARNING: Type "hge" (128 bit) is unsupported by Numpy. The
numbers are instead converted to python objects of type "PyLong". This means a
python object is constructed for every huge integer and the entire column is
copied.
+!PERFORMANCE WARNING: Type "hge" (128 bit) is unsupported by Numpy. The
numbers are instead converted to python objects of type "PyLong". This means a
python object is constructed for every huge integer and the entire column is
copied.
# 00:29:47 >
# 00:29:47 > "mclient" "-lmal" "-ftest" "-Eutf-8" "-i" "-e"
"--host=/var/tmp/mtest-30800" "--port=32717"
diff --git a/monetdb5/extras/pyapi/Tests/pyapi_types_numeric.stable.err
b/monetdb5/extras/pyapi/Tests/pyapi_types_numeric.stable.err
--- a/monetdb5/extras/pyapi/Tests/pyapi_types_numeric.stable.err
+++ b/monetdb5/extras/pyapi/Tests/pyapi_types_numeric.stable.err
@@ -26,7 +26,7 @@ stderr of test 'pyapi_types_numeric` in
# cmdline opt embedded_r = true
# cmdline opt embedded_py = true
# cmdline opt gdk_debug = 536870922
-!WARNING: Type "hge" (128 bit) is unsupported by Numpy. The numbers are
instead converted to python objects of type "long". This is likely very slow.
+!PERFORMANCE WARNING: Type "hge" (128 bit) is unsupported by Numpy. The
numbers are instead converted to python objects of type "PyLong". This means a
python object is constructed for every huge integer and the entire column is
copied.
# 19:41:50 >
# 19:41:50 > "mclient" "-lmal" "-ftest" "-Eutf-8" "-i" "-e"
"--host=/var/tmp/mtest-7261" "--port=31934"
diff --git a/monetdb5/extras/pyapi/bytearray.c
b/monetdb5/extras/pyapi/bytearray.c
new file mode 100644
--- /dev/null
+++ b/monetdb5/extras/pyapi/bytearray.c
@@ -0,0 +1,134 @@
+
+#include "bytearray.h"
+
+
_______________________________________________
checkin-list mailing list
checkin-list@monetdb.org
https://www.monetdb.org/mailman/listinfo/checkin-list
