Py-mdbm
- Py-mdbm is a Python binds to Yahoo! MDBM C API.
- MDBM is a super-fast memory-mapped key/value store.
- MDBM is an ndbm work-alike hashed database library based on sdbm which is based on Per-Aake Larson’s Dynamic Hashing algorithms.
- MDBM is a high-performance, memory-mapped hash database similar to the homegrown libhash.
- The records stored in a mdbm database may have keys and values of arbitrary and variable lengths.
| Build Stats |
Py-mdbm ver. |
License. |
Y! mdbm ver. |
 |
 |
 |
 |
Table of Contents
API
Currently Supported APIs
the following is list of support api on now.
| Group |
API |
| File Management |
mdbm_open, mdbm_close, mdbm_sync, mdbm_fsync, mdbm_close_fd, mdbm_replace_db, mdbm_replace_file, mdbm_dup_handle, mdbm_pre_split, mdbm_fcopy |
| Configuration |
mdbm_get_version, mdbm_get_size, mdbm_get_page_size, mdbm_get_limit_size, mdbm_get_hash, mdbm_get_alignment, mdbm_set_alignment, mdbm_setspillsize, mdbm_limit_dir_size, mdbm_get_magic_number, mdbm_limit_size_v3, mdbm_set_window_size |
| Record Access |
mdbm_fetch, mdbm_delete, mdbm_store, mdbm_fetch_r, mdbm_fetch_dup_r, mdbm_delete_r, mdbm_store_r, mdbm_fetch_info |
| Record Iteration |
mdbm_first, mdbm_next, mdbm_firstkey, mdbm_nextkey, mdbm_first_r, mdbm_next_r, mdbm_firstkey_r, mdbm_nextkey_r, mdbm_iterate |
| Locking |
mdbm_islocked, mdbm_isowned, mdbm_lock, mdbm_unlock, mdbm_lock_reset, mdbm_delete_lockfiles, mdbm_get_lockmode, mdbm_trylock, mdbm_plock, mdbm_punlock, mdbm_tryplock, mdbm_lock_shared, mdbm_trylock_shared, mdbm_lock_smart, mdbm_trylock_smart, mdbm_unlock_smart |
| Data Management |
mdbm_compress_tree, mdbm_truncate, mdbm_purge, mdbm_clean, mdbm_prune, mdbm_set_cleanfunc |
| Data Integrity |
mdbm_check, mdbm_chk_all_page, mdbm_chk_page, mdbm_protect |
| Data Display |
mdbm_dump_all_page, mdbm_dump_page |
| Statistics |
mdbm_count_records, mdbm_count_pages, mdbm_get_stats, mdbm_get_db_info, mdbm_get_stat_counter, mdbm_get_stat_time, mdbm_reset_stat_operations, mdbm_enable_stat_operations, mdbm_set_stat_time_func, mdbm_get_db_stats, mdbm_get_window_stats, mdbm_get_stat_name, mdbm_set_stats_func, mdbm_chunk_iterate |
| Cache and Backing Store |
mdbm_set_cachemode, mdbm_get_cachemode, mdbm_get_cachemode_name, mdbm_set_backingstore |
| Import and Export |
mdbm_cdbdump_to_file, mdbm_cdbdump_trailer_and_close, mdbm_cdbdump_add_record, mdbm_dbdump_to_file, mdbm_dbdump_trailer_and_close, mdbm_dbdump_add_record, mdbm_dbdump_export_header, mdbm_dbdump_import_header, mdbm_dbdump_import, mdbm_cdbdump_import |
| Miscellaneous |
mdbm_preload, mdbm_get_errno, mdbm_get_page, mdbm_lock_pages, mdbm_unlock_pages, mdbm_get_hash_value, mdbm_select_log_plugin, mdbm_set_log_filename |
Deprecated APIs
| API |
STATUS |
COMMENT |
| mdbm_save |
DEPRECATED |
mdbm_save is only supported for V2 MDBMs. |
| mdbm_restore |
DEPRECATED |
mdbm_restore is only supported for V2 MDBMs. |
| mdbm_sethash |
DEPRECATED |
Legacy version of mdbm_set_hash() This function has inconsistent naming, an error return value. It will be removed in a future version. |
Only a V2 implementation
| API |
STATUS |
COMMENT |
| mdbm_stat_all_page |
V3 not supported |
There is only a V2 implementation. V3 not currently supported. |
| mdbm_stat_header |
V3 not supported |
There is only a V2 implementation. V3 not currently supported. |
Has not been implemented
| API |
STATUS |
COMMENT |
| dbm_chk_error |
Not Implemented |
This has not been implemented |
Support two compatibility version
Python
| Version |
Support |
Test |
| Python 2.6.x ~ 2.7.x |
yes |
always |
| Python 3.0.x ~ 3.x.x |
yes |
always |
| PyPy |
yes |
always |
| PyPy3 |
yes |
always |
MDBM
| branch or release ver. |
Support |
Test |
Comment |
| master |
yes |
always |
|
| 4.x |
yes |
always |
|
Install
MDBM
py-mdbm (use pip)
pip install py-mdbm
py-mdbm (use source)
Download
git clone https://github.com/torden/py-mdbm
Build and Test
cd py-mdbm
CMD_PYTHON=`which python` make
Check
$ python
>>> import mdbm
>>> help(mdbm)
Benchmark
cd py-mdbm
`which pip` install -r for-benchmark-py26_or_higher-requirements.txt
CMD_PYTHON=`which python` CMD_PYTEST=`which pytest` make benchmark
Example
See the Source Code for more details
The following is Sample codes for a first look at py-mdbm
Creating and populating a database
Python 2 or higher
import mdbm
import random
print("[*] Creating and populating a database")
path = "/tmp/test1.mdbm"
flags = mdbm.MDBM_O_RDWR
flags = flags | mdbm.MDBM_O_CREAT
flags = flags | mdbm.MDBM_LARGE_OBJECTS
flags = flags | mdbm.MDBM_ANY_LOCKS
flags = flags | mdbm.MDBM_O_TRUNC
mode = 0o644 # means 0644
dbm = mdbm.open(path, flags, mode)
for i in range(0, 65535):
k = str(i)
v = str(random.randrange(0, 65535))
rv = dbm.store(k, v, mdbm.MDBM_INSERT)
if not rv:
print("[-] failed to data store to ", path)
break
print("[*] count of records : %d" % dbm.count_records())
dbm.close()
print("done")
Python 3 or higher
# encoding: utf-8
import mdbm
import random
print("[*] Creating and populating a database")
path = "/tmp/test1-byte.mdbm"
flags = mdbm.MDBM_O_RDWR
flags = flags | mdbm.MDBM_O_CREAT
flags = flags | mdbm.MDBM_LARGE_OBJECTS
flags = flags | mdbm.MDBM_ANY_LOCKS
flags = flags | mdbm.MDBM_O_TRUNC
mode = 0o644 # means 0644
dbm = mdbm.open(path, flags, mode)
print("|--------|-------|")
print("| key | val |")
print("|--------|-------|")
# byte
for i in range(0, 10):
k = bytes(str(i), 'utf-8')
v = bytes(str(random.randrange(0, 65535)), 'utf-8')
print("|%08s|%08s|" % (k, v))
rv = dbm.store(k, v, mdbm.MDBM_INSERT|mdbm.MDBM_CACHE_MODIFY)
if not rv:
print("[-] failed to data store to ", path)
break
# string
for i in range(10, 20):
k = str(i)
v = str(random.randrange(0, 65535))
print("|%08s|%08s|" % (k, v))
rv = dbm.store(k, v, mdbm.MDBM_INSERT|mdbm.MDBM_CACHE_MODIFY)
if not rv:
print("[-] failed to data store to ", path)
break
print("|--------|--------|")
print("[*] count of records : %d" % dbm.count_records())
print("\n")
dbm.close()
Fetching records in-place
import mdbm
import random
print("[*] Fetching records in-place")
path = "/tmp/test1.mdbm"
flags = mdbm.MDBM_O_RDWR
mode = 0o644 # means 0644
dbm = mdbm.open(path, flags, mode)
dbm.preload()
print("|-------|-------|")
print("| key | val |")
print("|-------|-------|")
for i in range(0, 10):
k = str(random.randrange(0, 65534))
orgval = dbm.fetch(k)
if not orgval:
print("[-] failed to fetch value of %s in mdbm" % k)
break
print("|%07s|%07s|" % (k, orgval))
print("|-------|-------|")
print("[*] count of records : %d" % dbm.count_records())
dbm.close()
print("done")
Fetching and updating records in-place
import mdbm
import random
print("[*] Fetching and updating records in-place")
path = "/tmp/test1.mdbm"
flags = mdbm.MDBM_O_RDWR
mode = 0o644 # means 0644
dbm = mdbm.open(path, flags, mode)
for i in range(0, 65535):
k = str(i)
v = str(random.randrange(0, 65535))
orgval = dbm.fetch(k)
if not orgval:
print("[-] failed to fetch value of %s in mdbm" % k)
break
print("[=] key(%s) : replace val(%s) to '%s' : " % (k, orgval, v)),
rv = dbm.store(k, v, mdbm.MDBM_REPLACE)
if not rv:
print("FAIL")
break
print("DONE")
print("[*] count of records : %d" % dbm.count_records())
dbm.close()
print("done")
Deleting records in-place
import mdbm
import random
print("[*] Deleting records in-place")
path = "/tmp/test1.mdbm"
flags = mdbm.MDBM_O_RDWR
mode = 0o644 # means 0644
dbm = mdbm.open(path, flags, mode)
for i in range(0, 10):
k = str(random.randrange(0, 65534))
rv = dbm.delete(k)
if not rv:
print("[-] failed to delete an record, key=%s" % k)
v = dbm.fetch(k)
if v:
print("[-] failed to delete an record, key=%s, val=%s" % (k,v))
break
print("[*] count of records : %d" % dbm.count_records())
dbm.close()
print("done")
Iterating over all records
import mdbm
import random
print("[*] Iterating over all records")
path = "/tmp/test1.mdbm"
flags = mdbm.MDBM_O_RDWR
mode = 0o644 # means 0644
dbm = mdbm.open(path, flags, mode)
print("|-------|-------|")
print("| key | val |")
print("|-------|-------|")
kv = dbm.first()
print("|%07s|%07s|" % kv)
while kv:
print("|%07s|%07s|" % kv)
kv = dbm.next()
print("|-------|-------|")
print("[*] count of records : %d" % dbm.count_records())
dbm.close()
print("done")
Iterating over all keys
import mdbm
import random
print("[*] Iterating over all records")
path = "/tmp/test1.mdbm"
flags = mdbm.MDBM_O_RDWR
mode = 0o644 # means 0644
dbm = mdbm.open(path, flags, mode)
print("|-------|")
print("| key |")
print("|-------|")
k = dbm.firstkey()
print("|%07s|" % k)
while k:
print("|%07s|" % k)
k = dbm.nextkey()
print("|-------|")
print("[*] count of records : %d" % dbm.count_records())
dbm.close()
print("done")
Iteration over all value by key
import mdbm
import random
print("[*] Creating and populating a database")
path = "/tmp/test_py_dup.mdbm"
flags = mdbm.MDBM_O_RDWR
flags = flags | mdbm.MDBM_O_CREAT
flags = flags | mdbm.MDBM_LARGE_OBJECTS
flags = flags | mdbm.MDBM_ANY_LOCKS
flags = flags | mdbm.MDBM_O_TRUNC
mode = 0o644 # means 0644
with mdbm.open(path, flags, mode) as dbm:
for k in range(0, 100):
key = str(k)
for i in range(1, 12):
val = str(123 * i)
rv = dbm.store(key, val, mdbm.MDBM_INSERT_DUP)
if not rv:
print("[-] failed to data store to ", path)
break
print("[*] Loop through DB, looking at records with the same key.")
with mdbm.open(path, mdbm.MDBM_O_RDONLY, mode) as dbm:
print("[*] count of records : %d" % dbm.count_records())
print("|-------|-------|")
print("| key | val |")
print("|-------|-------|")
k = str(random.randrange(0, 99))
empty_iter = dbm.init_iter()
info = dbm.fetch_dup_r(k, empty_iter)
while info:
print("|%07s|%07s|" % (k, info['val']))
info = dbm.fetch_dup_r(k, info['iter'])
print("|-------|-------|")
print("done")
Benchmark
The following is results of Py-mdbm vs AnyDBM vs SQLite3 vs Kyotocabinet benchmarks for simple data storing and random fetching in them.
Spec
Host
| Type |
Spec |
| CPU |
Inte i-7 |
| RAM |
DDR4 32G |
| HDD |
Nvme M.2 SSD |
VM
| Type |
Spec |
| Machine |
VM(VirtualBox) |
| OS |
CentOS 7 64bit |
| CPU |
2 vCore |
| RAM |
8G |
| AnyDBM |
Berkeley DB (Hash, version 9, native byte-order) format |
| Mdbm |
893f7a8 on 26 Jul, MDBM V3 format |
| SQLite |
V3 |
| Kyotocabinet |
1.2.76, kch |
Command
CMD_PYTHON=`which python` CMD_PYTEST=`which pytest` make benchmark
File Size
| Count of Records |
Type |
File Name |
Size |
| 10,000 |
SQLite3 |
test_py_benchmark_10000.db |
300K |
|
AnyDBM |
test_py_benchmark_10000.dbm |
348K |
|
Kyotocabinet KCH |
test_py_benchmark_10000.kch |
6.3M |
|
MDBM |
test_py_benchmark_10000.mdbm |
260K |
|
MDBM(TSC) |
test_py_benchmark_tsc_10000.mdbm |
260K |
| 100,000 |
SQLite3 |
test_py_benchmark_100000.db |
3.3M |
|
AnyDBM |
test_py_benchmark_100000.dbm |
2.5M |
|
Kyotocabinet KCH |
test_py_benchmark_100000.kch |
9.1M |
|
MDBM |
test_py_benchmark_100000.mdbm |
4.0M |
|
MDBM(TSC) |
test_py_benchmark_tsc_100000.mdbm |
4.0M |
| 1,000,000 |
SQLite3 |
test_py_benchmark_1000000.db |
35M |
|
AnyDBM |
test_py_benchmark_1000000.dbm |
39M |
|
Kyotocabinet KCH |
test_py_benchmark_1000000.kch |
37M |
|
MDBM |
test_py_benchmark_1000000.mdbm |
32M |
|
MDBM(TSC) |
test_py_benchmark_tsc_1000000.mdbm |
32M |
10,000 INSERTs
platform linux2 -- Python 2.7.14, pytest-3.3.2, py-1.5.2, pluggy-0.6.0
benchmark: 3.1.1 (defaults: timer=time.time disable_gc=False min_rounds=5 min_time=0.000005 max_time=1.0 calibration_precision=10 warmup=False warmup_iterations=100000)
rootdir: /root/PERSONAL/py-mdbm, inifile:
plugins: benchmark-3.1.1
collected 31 items
------------------------------------------------------------------------------------------- benchmark: 5 tests ------------------------------------------------------------------------------------------
Name (time in ms) Min Max Mean StdDev Median IQR Outliers OPS Rounds Iterations
---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
test_mdbm_store_tsc_10000 42.7790 (1.0) 46.4041 (1.0) 44.4735 (1.0) 0.8599 (1.0) 44.7228 (1.01) 1.0532 (1.0) 5;0 22.4853 (1.0) 23 1
test_mdbm_store_10000 43.0260 (1.01) 55.0859 (1.19) 45.1026 (1.01) 2.8206 (3.28) 44.1189 (1.0) 1.9995 (1.90) 3;2 22.1716 (0.99) 23 1
test_kyotocabinet_kch_store_10000 64.2769 (1.50) 72.2461 (1.56) 66.6182 (1.50) 2.1470 (2.50) 66.5540 (1.51) 2.4997 (2.37) 6;1 15.0109 (0.67) 16 1
test_sqlite3_store_10000 71.1770 (1.66) 89.0980 (1.92) 74.6003 (1.68) 4.5800 (5.33) 73.3149 (1.66) 2.8142 (2.67) 1;1 13.4048 (0.60) 13 1
test_anydbm_store_10000 129.4661 (3.03) 132.9770 (2.87) 131.7690 (2.96) 1.3268 (1.54) 132.4065 (3.00) 2.1240 (2.02) 1;0 7.5890 (0.34) 8 1
---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
100,000 INSERTs
------------------------------------------------------------------------------------------------ benchmark: 5 tests -----------------------------------------------------------------------------------------------
Name (time in ms) Min Max Mean StdDev Median IQR Outliers OPS Rounds Iterations
-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
test_mdbm_store_100000 432.5280 (1.0) 444.3109 (1.0) 440.1428 (1.0) 5.1283 (1.0) 443.0151 (1.0) 7.8554 (1.46) 1;0 2.2720 (1.0) 5 1
test_mdbm_store_tsc_100000 443.6021 (1.03) 457.2010 (1.03) 450.7210 (1.02) 6.5694 (1.28) 453.4068 (1.02) 12.3150 (2.28) 2;0 2.2187 (0.98) 5 1
test_kyotocabinet_kch_store_100000 553.1771 (1.28) 572.2950 (1.29) 559.4640 (1.27) 7.3967 (1.44) 557.5171 (1.26) 5.3908 (1.0) 1;1 1.7874 (0.79) 5 1
test_sqlite3_store_100000 668.3731 (1.55) 690.7680 (1.55) 676.8432 (1.54) 10.4372 (2.04) 670.3589 (1.51) 17.5762 (3.26) 1;0 1.4774 (0.65) 5 1
test_anydbm_store_100000 1,746.3379 (4.04) 1,778.0671 (4.00) 1,759.8858 (4.00) 12.6857 (2.47) 1,761.1270 (3.98) 19.0974 (3.54) 2;0 0.5682 (0.25) 5 1
-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
1,000,000 INSERTs
----------------------------------------------------------------------------------------- benchmark: 5 tests -----------------------------------------------------------------------------------------
Name (time in s) Min Max Mean StdDev Median IQR Outliers OPS Rounds Iterations
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
test_mdbm_store_1000000 4.4507 (1.0) 4.5549 (1.0) 4.5087 (1.0) 0.0386 (1.41) 4.5170 (1.00) 0.0471 (1.0) 2;0 0.2218 (1.0) 5 1
test_mdbm_store_tsc_1000000 4.4964 (1.01) 4.5557 (1.00) 4.5252 (1.00) 0.0275 (1.0) 4.5133 (1.0) 0.0494 (1.05) 3;0 0.2210 (1.00) 5 1
test_kyotocabinet_kch_store_1000000 5.5518 (1.25) 7.3104 (1.60) 5.9554 (1.32) 0.7585 (27.62) 5.6386 (1.25) 0.4548 (9.65) 1;1 0.1679 (0.76) 5 1
test_sqlite3_store_1000000 6.9506 (1.56) 7.1580 (1.57) 7.0168 (1.56) 0.0811 (2.95) 6.9938 (1.55) 0.0623 (1.32) 1;1 0.1425 (0.64) 5 1
test_anydbm_store_1000000 18.8494 (4.24) 19.3685 (4.25) 19.1384 (4.24) 0.1884 (6.86) 19.1481 (4.24) 0.1982 (4.21) 2;0 0.0523 (0.24) 5 1
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
10,000 Random Key SELECTs
----------------------------------------------------------------------------------------------- benchmark: 6 tests -----------------------------------------------------------------------------------------------
Name (time in ms) Min Max Mean StdDev Median IQR Outliers OPS Rounds Iterations
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
test_mdbm_random_fetch_10000 33.6039 (1.0) 37.1680 (1.02) 35.4372 (1.0) 0.8726 (1.63) 35.3181 (1.0) 1.0861 (1.73) 9;0 28.2189 (1.0) 29 1
test_mdbm_preload_random_fetch_tsc_10000 34.1651 (1.02) 36.5930 (1.0) 35.5276 (1.00) 0.5728 (1.07) 35.6691 (1.01) 0.6691 (1.06) 8;0 28.1471 (1.00) 29 1
test_mdbm_preload_random_fetch_10000 34.8370 (1.04) 37.1509 (1.02) 35.6486 (1.01) 0.5368 (1.0) 35.6290 (1.01) 0.6291 (1.0) 8;1 28.0516 (0.99) 27 1
test_kyotocabinet_random_fetch_10000 50.1349 (1.49) 315.4690 (8.62) 66.3761 (1.87) 60.3302 (112.39) 52.3400 (1.48) 1.3785 (2.19) 1;1 15.0657 (0.53) 19 1
test_anydbm_random_fetch_10000 98.3920 (2.93) 127.4319 (3.48) 103.2393 (2.91) 8.6436 (16.10) 101.2516 (2.87) 3.1178 (4.96) 1;1 9.6862 (0.34) 10 1
test_sqlite3_random_fetch_10000 179.9428 (5.35) 264.3309 (7.22) 198.3913 (5.60) 32.8237 (61.15) 183.5115 (5.20) 14.0412 (22.32) 1;1 5.0405 (0.18) 6 1
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
100,000 Random Key SELECTs
-------------------------------------------------------------------------------------------------- benchmark: 5 tests --------------------------------------------------------------------------------------------------
Name (time in ms) Min Max Mean StdDev Median IQR Outliers OPS Rounds Iterations
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
test_mdbm_preload_random_fetch_tsc_100000 351.3479 (1.0) 362.6180 (1.02) 358.0612 (1.01) 4.9707 (3.89) 360.7321 (1.02) 8.1980 (3.88) 1;0 2.7928 (0.99) 5 1
test_mdbm_preload_random_fetch_100000 352.9408 (1.00) 360.9550 (1.01) 356.8196 (1.01) 3.2021 (2.51) 357.4481 (1.01) 5.0185 (2.38) 2;0 2.8025 (0.99) 5 1
test_mdbm_random_fetch_100000 353.4501 (1.01) 356.4832 (1.0) 354.6917 (1.0) 1.2767 (1.0) 354.3482 (1.0) 2.1121 (1.0) 1;0 2.8193 (1.0) 5 1
test_kyotocabinet_random_fetch_100000 513.2129 (1.46) 516.0379 (1.45) 514.8367 (1.45) 1.3007 (1.02) 515.3730 (1.45) 2.3472 (1.11) 1;0 1.9424 (0.69) 5 1
test_anydbm_random_fetch_100000 1,196.3558 (3.41) 1,217.2129 (3.41) 1,207.2943 (3.40) 7.5601 (5.92) 1,206.6510 (3.41) 8.3598 (3.96) 2;0 0.8283 (0.29) 5 1
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
1,000,000 Random Key SELECTs
--------------------------------------------------------------------------------------------- benchmark: 5 tests --------------------------------------------------------------------------------------------
Name (time in s) Min Max Mean StdDev Median IQR Outliers OPS Rounds Iterations
-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
test_mdbm_preload_random_fetch_tsc_1000000 3.6708 (1.0) 3.7161 (1.0) 3.7020 (1.0) 0.0194 (1.29) 3.7138 (1.00) 0.0255 (1.71) 1;0 0.2701 (1.0) 5 1
test_mdbm_preload_random_fetch_1000000 3.6781 (1.00) 3.7315 (1.00) 3.7045 (1.00) 0.0212 (1.41) 3.7021 (1.0) 0.0336 (2.25) 2;0 0.2699 (1.00) 5 1
test_mdbm_random_fetch_1000000 3.6957 (1.01) 3.7336 (1.00) 3.7079 (1.00) 0.0150 (1.0) 3.7054 (1.00) 0.0149 (1.0) 1;0 0.2697 (1.00) 5 1
test_kyotocabinet_random_fetch_1000000 5.2549 (1.43) 5.2865 (1.42) 5.2677 (1.42) 0.0151 (1.01) 5.2599 (1.42) 0.0273 (1.82) 1;0 0.1898 (0.70) 5 1
test_anydbm_random_fetch_1000000 12.3323 (3.36) 12.4784 (3.36) 12.4044 (3.35) 0.0586 (3.90) 12.3911 (3.35) 0.0927 (6.20) 2;0 0.0806 (0.30) 5 1
-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Link
Please feel free. I hope it is helpful for you
