reflexive 1.2.7__py3-none-any.whl → 2.0.0__py3-none-any.whl

reflexive/res_analysis.py

@@ -0,0 +1,217 @@
+import logging
+import os
+from functools import (partial)
+import pandas as pd
+from reflexive.service import AWS_service
+from reflexive.analysis_functions import (
+    _clean_text,
+    _whitespace_cleaner,
+    _upload_text,
+    _s3_text_uploader,
+    _analyse_text,
+    _comprehend_cer_analyser,
+    _cer_job_progress,
+    _download_from_s3,
+    _extract_save_results,
+    _analysis_to_dataframe,
+    _add_offsets,
+    _offset_cleaner,
+    _orphan_joiner,
+    _add_res_sequence,
+    _add_res_interactions,
+    _add_res_weights,
+    _add_res_adj_matrix,
+    _jaccard_similarity,
+    _cosine_similarity,
+    _date_string,
+    _create_local_dir   
+)
+from reflexive.display_functions import (
+    _create_displacy_ents,
+    _render_annotated_text,
+    _create_graph,
+    _draw_graph
+)
+
+class RES_analyser:
+    
+    aws_service:AWS_service
+    config:dict
+    
+    logger = logging.getLogger(__name__)
+
+    
+    def __init__(self,parameters,prefix="res",postfix=None,dir="/data/")->None:
+        
+        self.config = self._build_config(prefix,postfix,dir)
+        
+        try:
+            self.aws_service = AWS_service(parameters)
+            self.aws_service.connect()
+            self.aws_service.get_s3_client()
+            self.aws_service.get_comprehend_client()
+        except Exception as e:
+            self.logger.error("There was an error setting up the AWS service: %s",repr(e))
+        else:
+            self.logger.info("AWS service setup successfully")
+    
+    ############################################################       
+    # MAIN ANALYSIS METHODS
+    
+    # Import files for analysis
+    def import_files_to_df(self,subdir="data_source/"):
+        return pd.DataFrame.from_dict(self._open_files_in_subdir(subdir))
+      
+    # Preprocess the text
+    # Uses clean_text() and whitespace_cleaner()
+    # Accepts a df, Returns a df
+    
+    preprocess_text = partial(_clean_text,text_cleaner=_whitespace_cleaner)
+    
+    # Upload text to s3
+    # Uses _upload_text()
+    # Accepts a df with 'text' and returns a df with 'uploaded' column
+
+    def upload_text_to_s3(self,df): 
+        return _upload_text(df,uploader=_s3_text_uploader,res_analyser=self)
+    
+    # Initiate comprehend custom entity analysis
+    # Uses _analyse_text()
+    # Accepts a df with 'uploaded' and returns a df with 'analysed' column
+    
+    def analyse_text(self):
+        return _analyse_text(analyser=_comprehend_cer_analyser,res_analyser=self)
+    
+    # Monitor analysis process
+    def monitor_job_progress(self,status,tz="UTC",useprint=False):
+        if useprint:
+            output = print
+        else:
+            output = self.logger.info 
+        return _cer_job_progress(status,self.aws_service,tz,output)
+    
+    # Download and extract results
+    def results_download_save_extract(self,status):
+        local_file = _download_from_s3(self,status)
+        return _extract_save_results(self,local_file)
+    
+    # Add results to dataframe
+    def add_res_results(self,df,results):
+        return _analysis_to_dataframe(df,results)
+
+    # Get offsets from results and add to dataframe
+    process_offsets = partial(_add_offsets,offset_cleaner=_offset_cleaner,orphan_joiner=_orphan_joiner)
+    
+    # Add text_display_ents
+    def add_text_display(self,df):
+        df['text_display_ents'] = [_create_displacy_ents(r.doc_name,r.text,r.offsets_clean) for i,r in df.iterrows()]
+        return df
+
+    # Create adjacency matrix from offsets
+    def add_interactions(self,df):
+        #Get RE sequence
+        df = _add_res_sequence(df)
+        df = _add_res_interactions(df)
+        df = _add_res_weights(df)
+        df = _add_res_adj_matrix(df)
+        return df
+
+    # Graph Jaccard Similarity
+    def get_jaccard_similarity(self,g1,g2):
+        return _jaccard_similarity(g1,g2)
+    
+    def get_cosine_similarity(self,m1,m2):
+        return _cosine_similarity(m1,m2)
+    
+    ############################################################
+    # UTILITY METHODS
+    
+    # Create the config dict used by s3 and comprehend methods
+    def _build_config(self,prefix:str,postfix:str,dir:str)->dict[str,str]:
+        if not postfix:
+            postfix = _date_string()
+        return {"local_data_dir": _create_local_dir(dir,logger=self.logger),
+                "s3_source_dir":f"{prefix}_files_{postfix}",
+                "s3_target_dir":f"{prefix}_results_{postfix}"}
+        
+    # Import files
+    def _open_files_in_subdir(self,subdir):
+        file_path = os.path.join(os.getcwd(),subdir)
+        file_names = []
+        texts = []
+        for file_name in sorted(os.listdir(file_path)):
+            file_names.append(file_name.split('.')[0])
+            with open(os.path.join(file_path,file_name),'r') as fp:
+                texts.append(fp.read())
+        return {"doc_name":file_names,"text":texts}
+    
+    
+    
+    
+class RES_visualiser:
+    
+    config:dict
+    
+    logger = logging.getLogger(__name__)
+
+    
+    def __init__(self)->None:
+        return None
+        # self.config = self._build_config()
+        
+        # try:
+            
+        # except Exception as e:
+        #     self.logger.error("There was an error setting up: %s",repr(e))
+        # else:
+        #     self.logger.info("setup successfully")
+    
+    ############################################################       
+    # MAIN VISUALISATION METHODS
+    
+    def show_annotated_text(self,ents):
+        return _render_annotated_text(ents)
+    
+    def save_annotated_text(self,name,ents,subdir="data/"):
+        file_path = os.path.join(os.getcwd(),subdir)
+        with open(f"{file_path}{name}.html","w") as fp:
+            fp.write(_render_annotated_text(ents,inline=False))
+    
+    def create_res_graph(self,matrix=None,id=None):
+        return _create_graph(matrix,id)
+    
+    def show_graph(self,graph):
+        return _draw_graph(graph,True)
+    
+    def save_graph(self,graph):
+        return _draw_graph(graph,False)
+    
+    ############################################################
+    # UTILITY METHODS
+    
+    # def show_df_graphs(self,df,scale=10,inline=True) -> str:   
+    #     for am in df.res_adj_matrix:
+    #         if scale > 1:
+    #             sm = self._scale_adj_matrix(am,scale)
+    #         else:
+    #             sm = am
+    #         g = self.create_res_graph(sm)
+    #         _draw_graph(g,True)
+    #     return ""
+    
+    # def _scale_adj_matrix(self,adj_matrix,scale):
+    #     new_adj = []
+    #     for row in adj_matrix:
+    #         new_row = []
+    #         for c in row:
+    #             new_row.append(round(c*scale,1))
+    #         new_adj.append(new_row)
+    #     return new_adj
+    
+    # Create the config dict used by s3 and comprehend methods
+    # def _build_config()->dict[str,str]:
+    #     return {"":"", 
+    #             "":"",
+    #             "":""}
+        
+        

reflexive/service.py

@@ -0,0 +1,58 @@
+import boto3
+import logging
+
+class AWS_service:
+    
+    logger = logging.getLogger(__name__)
+    
+    aws_params = {"profile":"default",
+                  "s3_access_point":"",
+                  "s3_bucket_name":"",
+                  "comprehend_service_role_name":"",
+                  "reflexive_entity_name":"",
+                  "reflexive_entity_version":""
+                  }
+    
+    aws_session:boto3.Session = None
+    aws_account_number:str = ""
+    s3_client = None
+    comprehend_client = None
+    
+    def __init__(self,params:dict[str,str])-> None:
+        self.aws_params = params
+        return None
+    
+    def connect(self)->None:
+        try:
+            self.aws_session = boto3.Session(profile_name=self.aws_params['profile'])
+            self.aws_account_number = self.aws_session.client('sts').get_caller_identity().get('Account')
+        except Exception as e:
+            self.logger.error("Unable to create an AWS session: %s",repr(e))
+        else:
+            self.logger.info("AWS session created successfully")
+        
+    def get_s3_client(self)->None:
+        try:
+            self.s3_client = self.aws_session.client(service_name='s3')
+        except Exception as e:
+            self.logger.error("Unable to get S3 client: %s",repr(e))
+        else:
+            self.logger.info("AWS s3 client obtained successfully")
+        return None
+    
+    def get_comprehend_client(self)->None:
+        try:
+            self.comprehend_client = self.aws_session.client(service_name='comprehend')
+        except Exception as e:
+            self.logger.error("Unable to get comprehend client: %s",repr(e))
+        else:
+            self.logger.info("AWS comprehend client obtained successfully")
+        return None
+             
+    # def region(self):
+    #     return self.aws_session.region_name
+    
+    # def access_key(self):
+    #     return self.aws_session.get_credentials().access_key
+
+    

{reflexive-1.2.7.dist-info → reflexive-2.0.0.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: reflexive
-Version: 1.2.7
+Version: 2.0.0
 Summary: Supports AWS Reflexive Expressions Systems (RES) Analysis
 Project-URL: Repository, https://github.com/nlytx/reflexive.git
 Author-email: Andrew Gibson <andrew@nlytx.io>

reflexive-2.0.0.dist-info/RECORD

@@ -0,0 +1,9 @@
+reflexive/__init__.py,sha256=UkA6xvSu40CGYguUuiBqBMnPBkld0h0GR9UnZrptdZI,167
+reflexive/analysis_functions.py,sha256=cVTQtQnQigMJSCoQOtXI0_tyX49Re96Uz-ubx4UToUw,13966
+reflexive/display_functions.py,sha256=2LdkINOgmZfiV7nkW0x_IeimxW3J80YIOOnjX21-RJA,5699
+reflexive/res_analysis.py,sha256=jpb2Fh_jLEZmi4t6I6Q9nVE9LYRc_YJMGx12_co_69Y,6841
+reflexive/service.py,sha256=O0MX2BCHTSNG_eW6LHBN1FOjaNTCGgYgh7vsk58NNAk,1927
+reflexive-2.0.0.dist-info/METADATA,sha256=60ZmW9iHzBxs3BZeUm6KrsRq3LFwDQB6xGS_zuRaYoo,574
+reflexive-2.0.0.dist-info/WHEEL,sha256=qtCwoSJWgHk21S1Kb4ihdzI2rlJ1ZKaIurTj_ngOhyQ,87
+reflexive-2.0.0.dist-info/licenses/LICENSE,sha256=xx0jnfkXJvxRnG63LTGOxlggYnIysveWIZ6H3PNdCrQ,11357
+reflexive-2.0.0.dist-info/RECORD,,