plugin oncrawl v1.0.0

oncrawl/code-env/python/desc.json

@@ -0,0 +1,7 @@
+{
+  "pythonInterpreter": "PYTHON36",
+  "acceptedPythonInterpreters": ["PYTHON36"],
+  "forceConda": false,
+  "installCorePackages": true,
+  "installJupyterSupport": false
+}

oncrawl/code-env/python/spec/requirements.txt

@@ -0,0 +1,2 @@
+prison
+pendulum

oncrawl/custom-recipes/data_queries/recipe.json

@@ -0,0 +1,30 @@
+{
+    "meta": {
+        "label": "OnCrawl data queries",
+        "description": "<p>Export URLs or aggregations from crawls or log monitoring events.</p><p><a href='mailto:[email protected]'>Contact us to get your API key</a></p>",        
+        "icon": "icon-globe",
+        "iconColor": "sky"
+    },
+
+    "kind": "PYTHON",
+
+    "inputRoles": [],
+
+    "outputRoles": [
+        {
+            "name": "output",
+            "label": "output",
+            "description": "output",
+            "arity": "UNARY",
+            "required": true,
+            "acceptsDataset": true
+        }
+    ],
+
+    "paramsModule" : "oncrawl-data_queries.module",
+    "paramsPythonSetup" : "data_queries.py",
+    "paramsTemplate" : "data_queries.html",
+
+    "params": []
+
+}

oncrawl/custom-recipes/data_queries/recipe.py

@@ -0,0 +1,185 @@
+# -*- coding: utf-8 -*-
+import dataiku
+from dataiku.customrecipe import get_output_names_for_role, get_recipe_config
+import oncrawl as oc
+from oncrawl import oncrawlDataAPI as ocd
+from oncrawl import oncrawlProjectAPI as ocp
+
+output_names = get_output_names_for_role('output')
+output_datasets = [dataiku.Dataset(name) for name in output_names]
+output = output_datasets[0]
+
+#------------------------------config & vars
+config = get_recipe_config()
+
+#config checker to raise better error
+e = None
+if 'api_key' not in config.keys():
+    e = 'Please add your API key'
+
+if 'list_projects_id_name' not in config.keys() or len(config['list_projects_id_name'].keys()) == 0:
+    e = 'Your Oncrawl account seems to have no projects available. Please check with your Oncrawl account.'
+
+if 'list_configs_crawls' not in config.keys() or len(config['list_configs_crawls'].keys()) == 0 or 'list_crawls_project' not in config.keys() or len(config['list_crawls_project'].keys()) == 0:
+    e = 'Your Oncrawl account seems to have no crawls available. Please check the choosen project and date range with your Oncrawl account.' 
+
+if e is not None:    
+    raise Exception(e)    
+
+headers = {
+        'Content-Type': 'application/json',
+        'Accept': 'application/json',
+        'Authorization' : 'Bearer {}'.format(config['api_key'])
+}
+
+#list project ids
+p_ids = []
+
+#if getting all projects : rebuild an up to date ids list
+if config['projects_id'] == 'all':
+    try:
+        p_ids_uptodate = ocp.get_projects(config['api_key'])
+
+        for p in p_ids_uptodate:
+            config['list_projects_id_name'][p['id']] = p['name']
+            p_ids.append(p['id'])
+
+    except Exception as e:
+        raise Exception(p_ids_uptodate)
+else:
+    p_ids = [config['projects_id'].split(',')[0]]
+
+#--list ids to get data : according config['index'], ids are related to projects or crawls - each id represents a crawl when index = pages or links and a project when index = logs
+if config['index'] == 'logs':
+
+    ids = p_ids
+
+else:
+
+    if config['crawls_id'] not in  ['all', 'last']:
+        ids = [config['crawls_id']]
+
+    else:
+
+        #if getting all or last crawls : rebuild an up to date ids list
+        try:
+            dates = oc.build_date_range(config)
+            date_start_yyyy_mm_dd = dates['start']
+            date_end_yyyy_mm_dd = dates['end']
+
+            crawl_start_timestamp = oc.datestring_to_miltimestamp_with_tz(dates['start'])
+            crawl_end_timestamp = oc.datestring_to_miltimestamp_with_tz(dates['end'])
+
+            limit = None
+
+            c_ids_uptodate = ocp.get_live_crawls(projects_id=p_ids, config=config, timestamp_range={'start': crawl_start_timestamp, 'end': crawl_end_timestamp}, limit=limit)
+
+            ids = []
+            count_crawls_by_projects = []
+            for c in c_ids_uptodate:
+                if c['config_name'] == config['crawl_config']:
+                    if (config['crawls_id'] == 'last' and c['project_id'] not in count_crawls_by_projects) or config['crawls_id'] != 'last':
+                        count_crawls_by_projects.append(c['project_id'])
+                        ids.append(c['id'])
+
+        except Exception as e:
+            raise
+
+
+#------------------------------schema
+#fields not returned by oncrawl API
+metadata = {
+    'project_id': 'string',
+    'project_name': 'string',
+    'crawl_id': 'string',
+    'config_name': 'string',
+    'crawl_start_timestamp': 'bigint'
+    }
+
+metadata_fields = ocd.build_schema_from_metadata(config, metadata)
+
+schema = {
+    'dataset_schema': metadata_fields['schema'],
+    'dataset_schema_field_list': metadata_fields['list']
+}
+fields_to_request_by_ids = {}
+
+for i, id in enumerate(ids):
+
+    progress = '#{} {} {}/{}'.format(id, 'crawls' if config['index'] != 'logs' else 'projects', (i+1), len(ids))
+
+    #when aggregate data, all items have same schema
+    if config['data_action'] == 'aggs':
+
+        if i == 0:
+            f = ocd.build_schema_from_config(config=config)
+            schema['dataset_schema'] = schema['dataset_schema'] + f
+            print('############################\r\n############################\r\nBuil dataset schema with: ', progress)
+        else:
+            break        
+    else:
+
+        print('############################\r\n############################\r\nBuil dataset schema with: ', progress)
+        #when export data, for many reasons all items could not have same schema        
+        #return new fields to add to dataset schema and all fields to request for this item
+        f = ocd.build_schema_from_oncrawl(config=config, id=id, headers=headers, schema=schema)
+
+        if 'item_schema' not in f.keys() or len(f['item_schema']) == 0:
+            continue
+
+        schema['dataset_schema'] = schema['dataset_schema'] + f['dataset_schema']
+        schema['dataset_schema_field_list'] = schema['dataset_schema_field_list'] + f['dataset_schema_field_list']
+
+        fields_to_request_by_ids[id] = f['item_schema']
+
+output.write_schema(schema['dataset_schema'])
+
+#------------------------------data & writer
+total_count = 0
+with output.get_writer() as writer:
+
+    for i, id in enumerate(ids):
+
+        #this case happend when project has no log feature or unexpected ES issue
+        if config['data_action'] == 'export' and id not in fields_to_request_by_ids.keys():
+            continue
+
+        progress = '#{} {} {}/{}'.format(id, 'crawls' if config['index'] != 'logs' else 'projects', (i+1), len(ids))
+        print('############################\r\n############################\r\nGet data for: ', progress)
+
+        metadata_value = ocd.fill_metadata(config, id)
+        if config['data_action'] == 'export':
+            data = ocd.export(config_oql=config['oql'], fields=fields_to_request_by_ids[id], config_index=config['index'], id=id, headers=headers)
+        else:
+            data = ocd.aggs(config_oql=config['oql'], config_index=config['index'], id=id, headers=headers)
+
+        count_result = 0
+        try:
+            for json_line in data:
+
+                row = metadata_value + []
+
+                if config['data_action'] == 'export':
+                    #oncrawl export api send values not in the same order as schema...
+                    for field in schema['dataset_schema_field_list']:
+                        if field not in list(metadata.keys()):
+                            if field in list(json_line.keys()):
+                                if field in ['title', 'meta_description', 'h1'] and json_line[field] is not None:
+                                    row.append(json_line[field].encode(encoding = 'utf8', errors = 'replace'))
+                                else:
+                                    row.append(json_line[field])
+                            else:
+                                row.append(None)
+                else:
+                    row = row + list(json_line.values())
+
+                writer.write_row_array(row)
+                count_result += 1
+                print(progress, 'row: ',count_result)
+            print(progress, ': total row recorded: ', count_result, '\r\n############################\r\n############################')
+
+        except Exception as e:
+            raise Exception('{}'.format(e))
+
+        total_count += 1
+