traitement par lots ok

actes-princiers/conf/base/catalog.yml

@@ -1,8 +1,17 @@
-xmlreflector:
-  type: actesdataset.XMLHousesReflector
+# ________________________________________________________________________
+
+# reading raw bourbon dataset
+bourbon:
+  type: actesdataset.XMLDataSetCollection
   housename: bourbon
   folderpath: data/01_raw/houses/bourbon
 
+# writing bourbon xmlcontent document attribute 
+bourbon_content:
+  type: actesdataset.XMLDataSetCollection
+  housename: bourbon
+  folderpath: data/02_intermediate/houses/bourbon
+
 actors:
   type: pandas.CSVDataSet
   filepath: data/01_raw/csv/actors.csv

actes-princiers/conf/base/parameters.yml

@@ -1 +1 @@
-xlststylesheet: templates/xsl/actes_princiers.xsl
+xsltstylesheet: templates/xsl/actes_princiers.xsl

actes-princiers/src/actes_princiers/customcontext.py

@@ -1,21 +1,7 @@
-from pathlib import Path
-from typing import Dict
+#from typing import Dict
 
 from kedro.framework.context import KedroContext
 
-from kedro.pipeline import Pipeline, node, pipeline
-from actesdataset import XMLDataSet
-
-
-# FIXME : move in an utils.py module
-def tree(directory, relative_to=None):
-    "helper that returns a directory tree structure" 
-    trees = dict()
-    for path in sorted(directory.rglob("*.xml")):
-        trees[path.stem] = str(path.relative_to(relative_to))
-    return trees
-
-
 class ProjectContext(KedroContext):
     project_name = "actes princiers"
     project_version = "0.1"
@@ -32,42 +18,11 @@ class ProjectContext(KedroContext):
         houses = self.config_loader.get("houses*")
         return houses['raw_datapath']
 
-    def _get_catalog(self, *args, **kwargs):
-        "catalog loader entry point"
-        # loading yaml defined catalogs 
-        catalog = super()._get_catalog(*args, **kwargs)
-        # kedro.io.data_catalog.DataCatalog
-        # adding data sets
-        self.nodes_description = self._house_dataset_loader(catalog)
-        return catalog 
-
-    def _house_dataset_loader(self, catalog):
-        nodes_description = []
-        houses_datapath = self.get_houses_datapath()
-        data_root_path = Path.cwd() / houses_datapath
-        relative_to = Path.cwd()
-        for dataset_name, dataset_path in tree(data_root_path, relative_to=relative_to).items():
-            catalog.add(data_set_name=dataset_name, 
-                data_set=XMLDataSet(filepath=dataset_path),
-                replace=True)
-            # adding an output catalog entry
-            output_dataset_name = dataset_name + "_output"
-            # XXX : make better
-            output_dataset_path = Path(dataset_path.replace("01_raw", "02_intermediate"))
-            # let's create subfolders if they don't exist
-            output_dataset_dir = output_dataset_path.parent
-            output_dataset_dir.mkdir(parents=True, exist_ok=True)
-            catalog.add(data_set_name=output_dataset_name, 
-                data_set=XMLDataSet(filepath=output_dataset_path),
-                replace=True)
-            # prepare information for the next stage (the pipeline stage)
-            node_description = dict(
-                inputs=dataset_name,
-                outputs=output_dataset_name,
-                name=dataset_name)
-            nodes_description.append(node_description)
-        return nodes_description
-
-    def prepare_pipeline_creation(self):
-        return self.nodes_description
-
+#    def _get_catalog(self, *args, **kwargs):
+#        "catalog loader entry point"
+#        # loading yaml defined catalogs 
+#        catalog = super()._get_catalog(*args, **kwargs)
+#        # kedro.io.data_catalog.DataCatalog
+#        # adding data sets
+#        self.nodes_description = self._house_dataset_loader(catalog)
+#        return catalog 

actes-princiers/src/actes_princiers/pipelines/xml_processing/nodes.py

@@ -1,8 +1,39 @@
+import logging
+from pathlib import Path
+
 from lxml import etree
 
-def parse_xsl(source_doc, xlststylesheet):
+from actesdataset import XMLDataSet
+
+logger = logging.getLogger(__name__)
+
+
+def transform(source_doc, xlststylesheet):
     #<class 'lxml.etree._XSLTResultTree'>
     xslt_doc = etree.parse(xlststylesheet)
     xslt_transformer = etree.XSLT(xslt_doc)
     return str(xslt_transformer(source_doc))
 
+
+def parse_xml_collection(datasets, param):
+    # FIXME set signature
+    # datasets -> dict
+    # param -> str
+#    logger.info("~~~~~~~~~~~~~~~~~~~~~~~~~~~~~")
+#    logger.info(str(bourbon))    
+    output_datasets = dict()
+#    datasets = bourbon.get_datasets()
+    for dataset_filenamestem, dataset in datasets.items():
+        # manually loading the dataset
+        dataset._load()
+        output_source_doc = transform(dataset.get_source_doc(), param)
+        output_filepath = dataset.get_filepath().replace("01_raw", "02_intermediate")
+        output_xmldataset = XMLDataSet(output_filepath)
+        output_xmldataset.set_source_doc(output_source_doc)
+        output_datasets[dataset_filenamestem] = output_xmldataset
+        # let's create subfolders if they don't exist
+        output_filepath = Path(output_filepath)
+        output_xmldataset_dir = output_filepath.parent
+        output_xmldataset_dir.mkdir(parents=True, exist_ok=True)
+    return output_datasets
+

actes-princiers/src/actes_princiers/pipelines/xml_processing/pipeline.py

@@ -1,37 +1,18 @@
-import logging
-
 from kedro.pipeline import Pipeline, node, pipeline
 from kedro.framework.session import KedroSession
 
-from .nodes import parse_xsl
-
-logger = logging.getLogger(__name__)
+from .nodes import parse_xml_collection
 
-# we need the context here in order to have access to prepare_pipeline_creation()
-with KedroSession.create() as session:
-    context = session.load_context()
-    # important: **we have to call** the catalog as an attribute, 
-    # because it makes a call to the _get_catalog() of the context method
-    catalog = context.catalog
-#    logger.info("loading houses")
-#    logger.info("------------------ houses ---------------------")
-#    logger.info(str(context.get_houses()))
 
-def nodes_factory(nodes_description):
-    "nodes creation"
-    nodes = []
-    for node_description in nodes_description:
-        node_name = node_description['name']
-        # logger.info(f"building node: {node_name}...")
-        nodes.append(node(
-                func=parse_xsl,
-                inputs=[node_description['inputs'], "params:xlststylesheet"],
-                outputs=node_description['outputs'],
-                name=node_name,
-                tags="xsl",
-            ))
-    return nodes
+def create_pipeline(**kwargs) -> Pipeline:
+    return pipeline(
+        [
+            node(
+                func=parse_xml_collection,
+                inputs=["bourbon", "params:xsltstylesheet"],
+                outputs="bourbon_content",
+                name="bourbon_ds_collection",
+            ),
+        ]
+    )
 
-def create_pipeline(**kwargs):
-    "pipeline entry point needed by the global pipeline registry"
-    return pipeline(nodes_factory(context.prepare_pipeline_creation()))

actes-princiers/src/actesdataset.py

@@ -1,28 +1,54 @@
 import json
 from typing import Dict, Any
+from pathlib import Path
+import logging
 
 from lxml import etree
 
 from kedro.io import AbstractDataSet, DataSetError
 
+from kedro.framework.session import KedroSession
 
-# FIXME: supprimer l'héritage
-class XMLDataSet(AbstractDataSet):
+logger = logging.getLogger(__name__)
+
+# FIXME : ça n'est plus utile
+# we need the context here to have access to the config
+with KedroSession.create() as session:
+    context = session.load_context()
+    
+class XMLDataSet:
     "lxml.etree._ElementTree loader"
-    # FIXME set the typing signature !!!!
 
+    # FIXME set the typing signature
     def __init__(self, filepath: str):
         self._filepath = filepath
 
-    def _load(self):
-        source_doc = etree.parse(self._filepath)
+    def get_filepath(self):
+        return self._filepath
+        
+    def get_source_doc(self):
+        if hasattr(self, 'source_doc'):
+            return self.source_doc
+        else:
+            attr_error_msg = str(self._describe())
+            raise AttributeError(f"XMLDataSet bject {attr_error_msg} has no attribute named : 'source_doc'")
+
+    def set_source_doc(self, source_doc):
+        self.source_doc = source_doc
+
+    def _transform_source_doc(self):
         # remove namespace :
         query = "descendant-or-self::*[namespace-uri()!='']"
-        for element in source_doc.xpath(query):
+        for element in self.source_doc.xpath(query):
             #replace element name with its local name
             element.tag = etree.QName(element).localname
-        etree.cleanup_namespaces(source_doc)
-        return source_doc
+        etree.cleanup_namespaces(self.source_doc)
+        return self.source_doc
+
+    def _load(self):
+        self.source_doc = etree.parse(self._filepath)
+        self._transform_source_doc()
+        return self.source_doc
         
     def _save(self, data:str) -> None:
         with open(self._filepath, 'w') as fhandle:
@@ -31,8 +57,9 @@ class XMLDataSet(AbstractDataSet):
     def _describe(self) -> Dict[str, Any]:
         return dict(filepath=self._filepath)
 
-class XMLHousesReflector(AbstractDataSet):
-    """``XMLHousesReflector`` stores instances of ``XMLDataSet``
+
+class XMLDataSetCollection(AbstractDataSet):
+    """Stores instances of ``XMLDataSet``
     implementations to provide ``load`` and ``save`` capabilities.
     anywhere in the program. To use a ``DataCatalog``, you need to
     instantiate it with a file system folder path, it "reflects"
@@ -50,26 +77,40 @@ class XMLHousesReflector(AbstractDataSet):
         >>> io = XMLCatalogReflector(housename='bourbon', folderpath='/tmp/mydir', data_sets={'cars': cars})
 #    filepath, load_args=None, save_args=None):
     """
+    # FIXME set the typing signature
     def __init__(self, 
         housename: str,
-        folderpath: str,
-        data_sets: dict[str, XMLDataSet] = None):
+        folderpath: str):
         self._housename = housename
-        self._folderpath = folderpath
-        self._datasets = data_sets
-#        self.filepath = filepath
-
+        self._folderpath = Path(folderpath)
+#        self.house_path = Path(context.get_houses()[self._housename]['path'])
+#        print(self.house_path)
+
+    def get_datasets(self):
+        if hasattr(self, 'datasets'):
+            return self.datasets
+        else:
+            attr_error_msg = str(self._describe())
+            raise AttributeError(f"Object {attr_error_msg} has no attribute named : 'datasets'")
+            
+    # FIXME : set the signature
     def _load(self):
-        return "C'est chargé!"
-        
-    def _save(self):
-        raise NotImplementedError("Attention : dataset en lecture seule !")
-        
-    def _exists(self) -> bool:
-        return True
+        ":return: dict[str, XMLDataSet]"
+        self.datasets = dict()
+        for filepath in sorted(self._folderpath.glob("*.xml")):
+            self.datasets[filepath.stem] = XMLDataSet(
+                filepath=str(filepath))
+        return self.datasets
+
+    # FIXME : set the signature
+    def _save(self, datasets):
+        # faire une méthode save et pas _save
+        for stemfilename, dataset in datasets.items():
+            # FIXME XXX -> pas besoin refaire un get_source_doc !!!!!!
+            dataset._save(dataset.get_source_doc())
         
     def _describe(self):
-        return dict(name="my own dataset")        
+        return dict(name=self._housename, folderpath=self._folderpath)
         
 #    def load(self, name: str) -> Any:
 #        """Loads a registered data set.
@@ -114,7 +155,7 @@ class XMLHousesReflector(AbstractDataSet):
 ##            >>> io.save("cars", df)
 ##        """
 ##        dataset = self._get_dataset(name)
-###        self._logger.info("Saving data to '%s' (%s)...", name, type(dataset).__name__)
+###        self._print("Saving data to '%s' (%s)...", name, type(dataset).__name__)
 ##        dataset.save(data)
 
 #    def _describe(self) -> Dict[str, Any]: