From 4029857c75c11cac8a55278ea71f6e926e3029b3 Mon Sep 17 00:00:00 2001
From: Pratishtha Abrol <pratishthaabrol@gmail.com>
Date: Wed, 16 Jun 2021 13:22:19 +0530
Subject: [PATCH 1/3] defining input and output basics

---
 .../__pycache__/component.cpython-38.pyc      | Bin 0 -> 2783 bytes
 component/__pycache__/executor.cpython-38.pyc | Bin 0 -> 3404 bytes
 component/component.py                        |  63 ++++++++++--------
 component/component_test.py                   |  42 ++++++------
 component/executor.py                         |  57 ++++++++++------
 5 files changed, 93 insertions(+), 69 deletions(-)
 create mode 100644 component/__pycache__/component.cpython-38.pyc
 create mode 100644 component/__pycache__/executor.cpython-38.pyc

diff --git a/component/__pycache__/component.cpython-38.pyc b/component/__pycache__/component.cpython-38.pyc
new file mode 100644
index 0000000000000000000000000000000000000000..15f6a6b824dac08929b1de84d8a31ac518eed318
GIT binary patch
literal 2783
zcma)8OK%)S5T4i0&dxq;Z=5)hkZ6z)EDGxz2n(T@#U!BMSXo=3xeVj!-raHLmF^y}
zE$5U-IR)_pa!783f5KmwD<}Q}B0yEo?5^_=uvV99s;jHJUSIj3+jTYg^?&*GyPsAx
z?Kk{5|FhuZ8Z7ZM5UP0^)u|EdiQyTE>6wY;SvtzhxRKbNt#~VLCXVMQz5#s8YboB2
z-K6cc72gED<8>78#NDLl^%UO%e#Ki+yc@43Yu=jT+i~CPU)N}db{}Z8%j}~I_f2md
zBt5zUk`?A0T?EM`+8%3z)nKM;+Rky1<S`@JgaqUUi{p%Zlku36{V%>HVIgFekT6T~
zEM=)2y6*lg60gXKh``7R;Vhtn$jPv_D#(Px`6$f`NrIG;JP?9`#*AmhbVdT<W`%^!
zAtA~z*iKlGN+L5N82_Gu3cTx11ZUxeLIzw$>6GMI1WK%a`y}Hegu8LeefiB|@#wi1
zB#1Mx@k1nMB$G1+hhOp&9Ek(TD0og8D8@0#Qt)s7BsWMGm!n%sBh4f^WJDBso^d>U
zc%p(RS&$6f-=Q?LfmvENBOy!YRxYC~4dT+;XUB4&mkoK6GXYHRFc8dNgd8MYcFe*;
zX51G!3sDfxf;45ZU&tuNGkpk5N&`-PjL#$pB}!LEmD*mwgM>-OQFBMxg1gU4-6N5u
zm9^avAA0`RfRAgi#8*I==IK!32Gu>28q8u1YEtWw5$T@InowUY+IXZZ&P9&ej}5O)
zn@b63cBljO*DCwtaK@5gRB%XLmX5)@(~m~;G~p+~)Q+m`U_n*~KvJC`3*}+iW5;1!
zPzFJYBPsR=MrmP8ODC&i{Y>Z3y&7z|5HAUNh(b|zRCW47J*Dhc^;m!I4!Wh|`zeIa
z_e<CJlZ+NI@*Us5TLf{v)AW6sg}%?*V81kl<fXf}eS7=so&BBLV~*&-Te!7WH^z8*
zzg_uM+F$PO9qf<G=B<N$WS(jKO+1I>kKN5#maxqn+&(?s{G5q9GRrr!oTa|V3Ldh}
zDO7WDxLLWesGiN&aKrqBcfshpu*5nLN4IQ4bN{rOuI~P2H-P_JX}nSjHfZBd+yHXE
z6qw~TnEgZ*z*F6Gsm<C)9ksfw$5yUrS2SuqHXmveom#Z<pzEzZ)GW<ggF<MQy($Lt
zw4NsaE9I_Buo5@`6Ulit<xB{&ML<f(2V^@8SuO>sB65{fiAC&z9)*O~8=$>AFnJ#w
z;unzgfP{$X^B+2m%Y-HNfSj3BzvJ*rQ|+E9wI_h}Pjy|B#*rzlBj|@m`aMh9XZo2T
zn@^lG^QnGjL62DX8nPwbGwr@{)IMuqcOYs*0PNsLK;luBLg5r4AZ1<W^LQ+)99<@4
zn@`1-;z)(9b-8YlZ9=0^AqzhNPXh@EmdYSX73j`=CR3I&D5VOcs@Pv==@RD=Yvts)
zzT$<J_?J8C*s4Y%!2|#YfVFN^z(Nds4YXBGmb;58t#8D(+9Gp|Bt;tCEf_)RknFM$
z@*IGF(S#&LEF&CW|794a2xql*LMRJk_@)rOWj*9k&SHqvP^s;nez>Cm1;<Sa(@3ET
z0DKq}f&u!U2O9!&3*lcdIpk|lA_MECisIO4fWMZ4@J`u`v*{E}msXSl#v#U*%^I&2
zRPZ-Iq`HuIaI-;+Bo}1^#tY;8EffJVMQJ2`e_(@M7EfMxCx07MpXp`G_rYf&0HhSG
z5D@eCfQ(=%7|_-M2z1xz8J2G7j&3}+t<#McJ9D_WQr{i)`9(Bx3CRYM%SaGfJa@de
zGxA3_x5s1NL;(O_HN!^u2!!_1nUA&7nQwaCoo{zW2m7~f`)ELgg};MVvH7@x1Y1)*
zyzJI4h45d?+%4tX`=~;YV4OaNC9uB{=PbR`(4HIMD7gAu?GHB8!%G{c1!SbUMcs#2
zl_oTTdWg3H+>nZRxF}aK#9_@59h9483sux4=MD-L=wDt4mnYRIuLgsf8S0=K=PKR&
yDqh`PG(HBbifs2YxN#v@4lL)yGPR}iNj0lJ#gQj4GnTt<baeQusNx)>-~R_uuNlDr

literal 0
HcmV?d00001

diff --git a/component/__pycache__/executor.cpython-38.pyc b/component/__pycache__/executor.cpython-38.pyc
new file mode 100644
index 0000000000000000000000000000000000000000..474a1949369446c2e91c9726695274f73664cfda
GIT binary patch
literal 3404
zcmb_fOK;rB5hgjD7iaXa{0<VN5oF;A*-$|aIT((yiY4!E;FSSOyBiF!peeFvhBC>f
zx|?fB7+ux|;v56)F$XEPMSe;C0<XCwe<48D`Kmb_O`OdkKsaEty1Tl%s=B_ao=;a+
zJPXfX|NNKdCu{gpXZdl_`2er_3QcT{tk{m7)E?R9>x>+HTWKqEM{d?0wX@EslXXX3
zoA}(cm-R<|L$^~e8;k~q?xe$PWwc`GZn~PSjn)j^OV_h&qidl1qm2)(*oz0Rt$4r(
zN7rAqMmJ(--`X68yS8QRo`hMEawf(sWFPS~73^~%)0iDR{w<42rA5Xfkrg86x%NHp
zV4A4^7-uShkW*HKN^!+_EmJW)m7<(XeP&i+VaCH;GcB0n@=MM_<%R4;lE<{Y(xKEz
zK4FDOz%&=a#zL}4^3Y7ZtB_h~)Tqe9R6xedL{FK}Q;x;&7{Me@L(OBD6>~7ADHA!A
zzfvh*Y#u^2!;uTk4mne0Q3y$k4^NGUG!HZ1`#Vn5+S_!i)^>hcxjRXuEB8sF&>rv;
zy=hl%eOhn@V&yPYJm4ohDz%Up=!}z;Cj!(e**^{=9h5priMf}Enm!w<Xv(wDFAnvI
zj%d5(r;W}Bc-2Q}xP_3iD4a$u?(#P8@a|FXt~GY<THo34(@gBdtv|Z48@FG(BhLg#
zr|Rv_^*sB2Lz(k1aq_5fxlfjRNnVsXh{^PUDD|SVMz#ip6a|NQTvZ*(bt&_R@X$O1
z5Iwvtyy|agX4Zu@vuDn%b=k5lZC`-8Xc1B_olCp;(Zu?_rQJ7{ZohKWL*0p6y8Fg@
zW#7X4m92Xh{TWbm*SfUNUG%(}do%#u{+%_p<96J6-J7*{z<b`&!;6(m>tYq(HE^zD
zU6<C;@4S1pcE|eSd*{7b?~V1MBmZ>#wZ3+-5u<<U*w*pav+j9cUq8BWans0fY8m?0
zrE~5<^6i;7>l3G0)t`Ap#~x5P2shw}N1`|dS~#0~G1MWUghm=)VhF*`@H_@0G{LZe
zh@OU;p&ufY4NdD72Nnyi{CZW$;$%GL5>Pgh-i5~dwOojL3FL6IKcJVgahXRt5qX#<
z`jo&*J1o+Ik@ZrP$Qr;%l1FJ7^LVapoiLhMVTS-E5E#Z#m%t?xjf^qRLh5~XSZXAq
zg_`edk|mR=Hp{P8vj#Sb6ss|B&^A|KTTaw{qathEiy*zvKBj=hPLXC~i7aPa&t7mG
zgn$Yu;A^tt9F)+lwhSq%5y#li8j(1j`)qr$iWFeak<#)wlyR_7(Qhn$4i_iU)MS^r
za&)$0ta%Qja~&o*(=U-J3-}zH-_sI9_l({5_?KaN&!4YrBE(qr)Ie{^tpC}dM%@3&
zqEeNJ%hM~H5IHK$&=|u|1D0IanDkE0AkUhE!U5A1@}tPvRZjUYN0Vcl`|N2^M=mS#
z<hbOFMt!!&Bd$~^PZ<!wvNF{PmBG3KE+Shk9Zd*8#5_vzjiIuEfH9;<3Z5o8pU2xX
zlin5~w};AE<MvsYpgK0O^K^e#N+6j~fCENpIfQJgpJvZ$m7f?K1+||lH!sul>^J{D
zN_u>kLs-r^5QaboAnRE6*@viM?~%6^s#fe*9>P;2%9BV%xCtyC2NWge>v+{2G!vxb
zFV@fPnT_PvmF~>`gLCZeTi;lccs8BN@qb>mjiTS!avfu4;vb)EO-07H3K{CNv#lMj
zUT9Hlp%UbQ5~Ymz)&zCBJlv|Q+rsTzS5$4CkbZGmc@GGbH1}|ybnBq$*bXYk*$;Q;
zmCD}{{)1FRVX7X2X!xs6y~WwD9+l9Bt{L?J3nzdTpx$J(lB)ZaXH2~_mce)a+0A`Z
zhfUrhq{71(-rKxW4TDrn0&|Hei^@AvBCl(A)lLP{UDcX!jWVfdw=j=*<&t)l_gR>h
z+{8lVQreJxm{nZ|cL3f=%7ev<4a6TVEp<o)6v(O?80G4Yq#I~8taW^bL*}w_DcRQI
z=$aFhGJ!x8YFVu=?SbK{3kxRLvT~DrEGkE+$}K`Yt=uDA$*PTDXZP{;r%w)o#~(l0
z4feJ_*{#-10uRi!B%qtj=Ad#FPsefyRpbgat7t~6<WjXrgrjS9e_8n-6Z2haZlbAH
z=8332r;#`BLQxZ3@)mL4CQfHZ`~+WhfX4d4YRk2~x32BHbsW!j?atedORej;wu{!o
zOvmkEhDO2l)@`rbJ=;Owv7K+h?R4rn$N8u04(+qSQgp$<USQlE1eF&AnTSiAQGfsl
zj*;r<RM6`NK`f#mkaVSybU~AZDM>L~^|t3lMH1HJFQ_3TEj47Aq-)W8cojJl8awOu
zy6sucw`OMZru-$waXN@K`BNf02p}9_E!OO8Hl^*WmL8wf+usJ|$13#~FjGEVOJ$e%
zyk<5LlI~<xp9V}ttwYoF%@x(eoAilk_-hDg++@^SGRD|uj;i;d_SZxFL$cHjG;Y5K
O=ig}!?fK7foqqv2tO<_*

literal 0
HcmV?d00001

diff --git a/component/component.py b/component/component.py
index ff9345a..7642809 100644
--- a/component/component.py
+++ b/component/component.py
@@ -22,57 +22,58 @@
 an example and will not be supported by TFX team.
 """
 
-from typing import Optional, Text
+from typing import List, Optional, Text
 
 from tfx import types
 from tfx.dsl.components.base import base_component
 from tfx.dsl.components.base import executor_spec
-from tfx.examples.custom_components.hello_world.hello_component import executor
+# from tfx.examples.custom_components.hello_world.hello_component import executor
 from tfx.types import channel_utils
 from tfx.types import standard_artifacts
 from tfx.types.component_spec import ChannelParameter
 from tfx.types.component_spec import ExecutionParameter
+from tfx.utils import json_utils
 
+import executor
 
-class HelloComponentSpec(types.ComponentSpec):
-  """ComponentSpec for Custom TFX Hello World Component."""
+class SchemaCurationSpec(types.ComponentSpec):
+  """ComponentSpec for Custom TFX Schema Curation Component."""
 
   PARAMETERS = {
       # These are parameters that will be passed in the call to
       # create an instance of this component.
-      'name': ExecutionParameter(type=Text),
+      # 'name': ExecutionParameter(type=Text),
+      'exclude_splits': ExecutionParameter(type=str, optional=True),
   }
   INPUTS = {
-      # This will be a dictionary with input artifacts, including URIs
-      'input_data': ChannelParameter(type=standard_artifacts.Examples),
+      'statistics': ChannelParameter(type=standard_artifacts.ExampleStatistics),
+      'input_schema': ChannelParameter(type=standard_artifacts.Schema),       # Dictionary obtained as output from SchemaGen
   }
   OUTPUTS = {
-      # This will be a dictionary which this component will populate
-      'output_data': ChannelParameter(type=standard_artifacts.Examples),
+      'output_schema': ChannelParameter(type=standard_artifacts.Schema),      # Dictionary which containes new schema
   }
 
 
-class HelloComponent(base_component.BaseComponent):
-  """Custom TFX Hello World Component.
+class SchemaCuration(base_component.BaseComponent):
+  """Custom TFX Schema Curation Component.
 
-  This custom component class consists of only a constructor.
+  Current progress : 
+    - Accepts schema, outputs the same schema
   """
 
-  SPEC_CLASS = HelloComponentSpec
+  SPEC_CLASS = SchemaCurationSpec
   EXECUTOR_SPEC = executor_spec.ExecutorClassSpec(executor.Executor)
 
   def __init__(self,
-               input_data: types.Channel = None,
-               output_data: types.Channel = None,
-               name: Optional[Text] = None):
-    """Construct a HelloComponent.
+               statistics: types.Channel,
+               input_schema: types.Channel,
+               exclude_splits: Optional[List[Text]] = None):
+    """Construct a SchemaCurationComponent.
 
     Args:
-      input_data: A Channel of type `standard_artifacts.Examples`. This will
-        often contain two splits: 'train', and 'eval'.
-      output_data: A Channel of type `standard_artifacts.Examples`. This will
-        usually contain the same splits as input_data.
-      name: Optional unique name. Necessary if multiple Hello components are
+      input_schema: A dictionary that containes the schema generated by SchemaGen component of tfx
+      output_schema: A dictionary that contains the schema after curation by the custom schema curation component
+      name: Optional unique name. Necessary if multiple custom schema curation components are
         declared in the same pipeline.
     """
     # output_data will contain a list of Channels for each split of the data,
@@ -80,9 +81,17 @@ def __init__(self,
     # passes the input data through to output, the splits in output_data will
     # be the same as the splits in input_data, which were generated by the
     # upstream component.
-    if not output_data:
-      output_data = channel_utils.as_channel([standard_artifacts.Examples()])
+    # if not output_schema:
+      # output_schema = channel_utils.as_channel([standard_artifacts.Examples()])
 
-    spec = HelloComponentSpec(input_data=input_data,
-                              output_data=output_data, name=name)
-    super(HelloComponent, self).__init__(spec=spec)
+    if exclude_splits is None:
+      exclude_splits = []
+      logging.info('Excluding no splits because exclude_splits is not set.')
+
+    output_schema = types.Channel(type=standard_artifacts.Schema)
+
+    spec = SchemaCurationSpec(statistics=statistics,
+                              input_schema=input_schema,
+                              exclude_splits=json_utils.dumps(exclude_splits),
+                              output_schema=output_schema)
+    super(SchemaCuration, self).__init__(spec=spec)
diff --git a/component/component_test.py b/component/component_test.py
index c77a48a..2ad3dd0 100644
--- a/component/component_test.py
+++ b/component/component_test.py
@@ -12,41 +12,37 @@
 # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 # See the License for the specific language governing permissions and
 # limitations under the License.
-"""Tests for HelloComponent."""
+"""Tests for SchemaCuration Component."""
 
 import json
 
 import tensorflow as tf
 
-from tfx.examples.custom_components.hello_world.hello_component import component
+# from tfx.examples.custom_components.hello_world.hello_component import component
 from tfx.types import artifact
 from tfx.types import channel_utils
 from tfx.types import standard_artifacts
+from tfx.types import standard_component_specs
+from tfx.types import artifact_utils
+import component
 
 
-class HelloComponentTest(tf.test.TestCase):
-
-  def setUp(self):
-    super(HelloComponentTest, self).setUp()
-    self.name = 'HelloWorld'
+class SchemaCurationTest(tf.test.TestCase):
 
   def testConstruct(self):
-    input_data = standard_artifacts.Examples()
-    input_data.split_names = json.dumps(artifact.DEFAULT_EXAMPLE_SPLITS)
-    output_data = standard_artifacts.Examples()
-    output_data.split_names = json.dumps(artifact.DEFAULT_EXAMPLE_SPLITS)
-    this_component = component.HelloComponent(
-        input_data=channel_utils.as_channel([input_data]),
-        output_data=channel_utils.as_channel([output_data]),
-        name=u'Testing123')
-    self.assertEqual(standard_artifacts.Examples.TYPE_NAME,
-                     this_component.outputs['output_data'].type_name)
-    artifact_collection = this_component.outputs['output_data'].get()
-    for artifacts in artifact_collection:
-      split_list = json.loads(artifacts.split_names)
-      self.assertEqual(artifact.DEFAULT_EXAMPLE_SPLITS.sort(),
-                       split_list.sort())
-
+    statistics_artifact = standard_artifacts.ExampleStatistics()
+    statistics_artifact.split_names = artifact_utils.encode_split_names(
+        ['train', 'eval'])
+    exclude_splits = []
+    schema_curation = component.SchemaCuration(
+        statistics=channel_utils.as_channel([statistics_artifact]),
+        input_schema=channel_utils.as_channel([standard_artifacts.Schema()]),
+        exclude_splits=exclude_splits,
+        )
+    self.assertEqual(
+        standard_artifacts.Schema.TYPE_NAME,
+        schema_curation.outputs['output_schema'].type_name)
+    
 
 if __name__ == '__main__':
   tf.test.main()
diff --git a/component/executor.py b/component/executor.py
index 808ff10..e0590c2 100644
--- a/component/executor.py
+++ b/component/executor.py
@@ -32,6 +32,7 @@
 from tfx.types import artifact_utils
 from tfx.utils import io_utils
 
+_DEFAULT_FILE_NAME = 'schema.pbtxt'
 
 class Executor(base_executor.BaseExecutor):
   """Executor for HelloComponent."""
@@ -67,22 +68,40 @@ def Do(self, input_dict: Dict[Text, List[types.Artifact]],
     """
     self._log_startup(input_dict, output_dict, exec_properties)
 
-    input_artifact = artifact_utils.get_single_instance(
-        input_dict['input_data'])
-    output_artifact = artifact_utils.get_single_instance(
-        output_dict['output_data'])
-    output_artifact.split_names = input_artifact.split_names
-
-    split_to_instance = {}
-
-    for split in json.loads(input_artifact.split_names):
-      uri = artifact_utils.get_split_uri([input_artifact], split)
-      split_to_instance[split] = uri
-
-    for split, instance in split_to_instance.items():
-      input_dir = instance
-      output_dir = artifact_utils.get_split_uri([output_artifact], split)
-      for filename in fileio.listdir(input_dir):
-        input_uri = os.path.join(input_dir, filename)
-        output_uri = os.path.join(output_dir, filename)
-        io_utils.copy_file(src=input_uri, dst=output_uri, overwrite=True)
+    # Load and deserialize exclude splits from execution properties.
+    exclude_splits = json_utils.loads(
+        exec_properties.get('exclude_splits', 'null')) or []
+    if not isinstance(exclude_splits, list):
+      raise ValueError('exclude_splits in execution properties needs to be a '
+                       'list. Got %s instead.' % type(exclude_splits))
+
+    # Setup output splits.
+    stats_artifact = artifact_utils.get_single_instance(
+        input_dict['statistics'])
+    stats_split_names = artifact_utils.decode_split_names(
+        stats_artifact.split_names)
+    split_names = [
+        split for split in stats_split_names if split not in exclude_splits
+    ]
+
+    schema = io_utils.SchemaReader().read(
+        io_utils.get_only_uri_in_dir(
+            artifact_utils.get_single_uri(
+                input_dict['schema'])))
+
+    for split in artifact_utils.decode_split_names(stats_artifact.split_names):
+      if split in exclude_splits:
+        continue
+      logging.info(
+          'Curating schema against the computed statistics for '
+          'split %s.', split)
+
+    output_schema = schema
+
+    output_uri = os.path.join(
+        artifact_utils.get_single_uri(
+            output_dict['schema']),
+        _DEFAULT_FILE_NAME)
+    io_utils.write_pbtxt_file(output_uri, output_schema)
+    logging.info('Schema written to %s.', output_uri)
+    
\ No newline at end of file

From 4d6f59c6dcb261f4d7f734ec1c897447be050c3a Mon Sep 17 00:00:00 2001
From: Pratishtha Abrol <pratishthaabrol@gmail.com>
Date: Wed, 16 Jun 2021 13:38:14 +0530
Subject: [PATCH 2/3] removing references to Hello-World Boilerplate

---
 component/component.py      | 12 ++----------
 component/component_test.py |  3 +--
 component/executor.py       | 10 ++--------
 3 files changed, 5 insertions(+), 20 deletions(-)

diff --git a/component/component.py b/component/component.py
index 7642809..2a2f0cc 100644
--- a/component/component.py
+++ b/component/component.py
@@ -12,14 +12,7 @@
 # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 # See the License for the specific language governing permissions and
 # limitations under the License.
-"""Example of a Hello World TFX custom component.
-
-This custom component simply reads tf.Examples from input and passes through as
-output.  This is meant to serve as a kind of starting point example for creating
-custom components.
-
-This component along with other custom component related code will only serve as
-an example and will not be supported by TFX team.
+"""TFX Schema Curation Custom Component
 """
 
 from typing import List, Optional, Text
@@ -27,7 +20,6 @@
 from tfx import types
 from tfx.dsl.components.base import base_component
 from tfx.dsl.components.base import executor_spec
-# from tfx.examples.custom_components.hello_world.hello_component import executor
 from tfx.types import channel_utils
 from tfx.types import standard_artifacts
 from tfx.types.component_spec import ChannelParameter
@@ -37,7 +29,7 @@
 import executor
 
 class SchemaCurationSpec(types.ComponentSpec):
-  """ComponentSpec for Custom TFX Schema Curation Component."""
+  """ComponentSpec for TFX Schema Curation Custom Component."""
 
   PARAMETERS = {
       # These are parameters that will be passed in the call to
diff --git a/component/component_test.py b/component/component_test.py
index 2ad3dd0..7524604 100644
--- a/component/component_test.py
+++ b/component/component_test.py
@@ -12,13 +12,12 @@
 # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 # See the License for the specific language governing permissions and
 # limitations under the License.
-"""Tests for SchemaCuration Component."""
+"""Tests for TFX Schema Curation Custom Component."""
 
 import json
 
 import tensorflow as tf
 
-# from tfx.examples.custom_components.hello_world.hello_component import component
 from tfx.types import artifact
 from tfx.types import channel_utils
 from tfx.types import standard_artifacts
diff --git a/component/executor.py b/component/executor.py
index e0590c2..7d14067 100644
--- a/component/executor.py
+++ b/component/executor.py
@@ -12,13 +12,7 @@
 # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 # See the License for the specific language governing permissions and
 # limitations under the License.
-"""Example of a Hello World TFX custom component.
-
-This custom component simply passes examples through. This is meant to serve as
-a kind of starting point example for creating custom components.
-
-This component along with other custom component related code will only serve as
-an example and will not be supported by TFX team.
+"""Executor code for TFX Schema Curation Custom Component
 """
 
 import json
@@ -35,7 +29,7 @@
 _DEFAULT_FILE_NAME = 'schema.pbtxt'
 
 class Executor(base_executor.BaseExecutor):
-  """Executor for HelloComponent."""
+  """Executor for TFX Schema Curation Custom Component."""
 
   def Do(self, input_dict: Dict[Text, List[types.Artifact]],
          output_dict: Dict[Text, List[types.Artifact]],

From 1108e43241ff0af85a507352e577c5b95e8b2554 Mon Sep 17 00:00:00 2001
From: Pratishtha Abrol <pratishthaabrol@gmail.com>
Date: Thu, 17 Jun 2021 16:33:34 +0530
Subject: [PATCH 3/3] module file parameters added to component spec class

---
 component/component.py | 12 +++++++++---
 1 file changed, 9 insertions(+), 3 deletions(-)

diff --git a/component/component.py b/component/component.py
index 2a2f0cc..eb0073a 100644
--- a/component/component.py
+++ b/component/component.py
@@ -32,9 +32,9 @@ class SchemaCurationSpec(types.ComponentSpec):
   """ComponentSpec for TFX Schema Curation Custom Component."""
 
   PARAMETERS = {
-      # These are parameters that will be passed in the call to
-      # create an instance of this component.
-      # 'name': ExecutionParameter(type=Text),
+      'module_file': ExecutionParameter(type=str, optional=True),
+      'module_path': ExecutionParameter(type=str, optional=True),
+      'preprocessing_fn': ExecutionParameter(type=str, optional=True),
       'exclude_splits': ExecutionParameter(type=str, optional=True),
   }
   INPUTS = {
@@ -49,6 +49,12 @@ class SchemaCurationSpec(types.ComponentSpec):
 class SchemaCuration(base_component.BaseComponent):
   """Custom TFX Schema Curation Component.
 
+  The SchemaCuration component is used to apply user code to a chema generated by SchemaGen
+  in order to curate the schema based on domain knowledge. 
+
+  Component `outputs` contains:
+   - `output_schema`: Channel of type `standard_artifact.Schema`
+
   Current progress : 
     - Accepts schema, outputs the same schema
   """