added document timestamps and doc source option for communication pro… (#112)

- Added timestamp that keeps track of when documents are accessed
- Added source function to communication protocol
- Added filter to each dashboard to select docs accessed at a specific date and time

Author: bradley-erickson

learning_observer/learning_observer/communication_protocol/executor.py

@@ -84,14 +84,14 @@ async def call_dispatch(functions, function_name, args, kwargs):
     >>> asyncio.run(call_dispatch({'double': double}, 'nonexistent', [1], {}))
     Traceback (most recent call last):
       ...
-    learning_observer.communication_protocol.exception.DAGExecutionException: ('Function nonexistent did not execute properly during call.', 'call_dispatch', {'function_name': 'nonexistent', 'args': [1], 'kwargs': {}, 'error': "'nonexistent'"})
+    learning_observer.communication_protocol.exception.DAGExecutionException: ('Function nonexistent did not execute properly during call.', 'call_dispatch', {'function_name': 'nonexistent', 'args': [1], 'kwargs': {}, 'error': "'nonexistent'"}, ...)
 
 
     Raises an exception when the called function raises an exception.
     >>> asyncio.run(call_dispatch({'double': double}, 'double', [None], {}))
     Traceback (most recent call last):
       ...
-    learning_observer.communication_protocol.exception.DAGExecutionException: ('Function double did not execute properly during call.', 'call_dispatch', {'function_name': 'double', 'args': [None], 'kwargs': {}, 'error': 'Input cannot be None'})
+    learning_observer.communication_protocol.exception.DAGExecutionException: ('Function double did not execute properly during call.', 'call_dispatch', {'function_name': 'double', 'args': [None], 'kwargs': {}, 'error': 'Input cannot be None'}, ...)
     """
     try:
         function = functions[function_name]
@@ -182,7 +182,7 @@ def handle_join(left, right, left_on, right_on):
     ...     right=[{'rid': 2, 'right': True}, {'rid': 1, 'right': True}],
     ...     left_on='lid', right_on='rid'
     ... )
-    [{'error': 'KeyError: key not found', 'function': 'handle_join', 'error_provenance': {'target': {'left': True}, 'key': 'lid', 'exception': KeyError("Key lid not found in {'left': True}")}, 'timestamp': ... 'traceback': ... {'lid': 2, 'left': True, 'rid': 2, 'right': True}]
+    [{'error': "KeyError: key `lid` not found in `dict_keys(['left'])`", 'function': 'handle_join', 'error_provenance': {'target': {'left': True}, 'key': 'lid', 'exception': KeyError("Key lid not found in {'left': True}")}, 'timestamp': ... 'traceback': ... {'lid': 2, 'left': True, 'rid': 2, 'right': True}]
     """
     right_dict = {}
     for d in right:
@@ -196,7 +196,6 @@ def handle_join(left, right, left_on, right_on):
     for left_dict in left:
         try:
             lookup_key = get_nested_dict_value(left_dict, left_on)
-
             right_dict_match = right_dict.get(lookup_key)
 
             if right_dict_match:
@@ -206,11 +205,12 @@ def handle_join(left, right, left_on, right_on):
                 merged_dict = left_dict
             result.append(merged_dict)
         except KeyError as e:
-            result.append(DAGExecutionException(
-                f'KeyError: key not found',
-                inspect.currentframe().f_code.co_name,
-                {'target': left_dict, 'key': left_on, 'exception': e}
-            ).to_dict())
+            result.append(left_dict)
+            # result.append(DAGExecutionException(
+            #     f'KeyError: key `{left_on}` not found in `{left_dict.keys()}`',
+            #     inspect.currentframe().f_code.co_name,
+            #     {'target': left_dict, 'key': left_on, 'exception': e}
+            # ).to_dict())
 
     return result
 
@@ -345,13 +345,13 @@ async def handle_map(functions, function_name, values, value_path, func_kwargs=N
     and handled later by the DAG executor. In our text, we return both a normal result
     and the result of an exception being caught.
     >>> asyncio.run(handle_map({'double': double}, 'double', [{'path': i} for i in [1, 'fail']], 'path'))
-    [{'output': 2, 'provenance': {'function': 'double', 'func_kwargs': {}, 'value': {'path': 1}, 'value_path': 'path'}}, {'error': 'Function double did not execute properly during map.', 'function': 'annotate_map_metadata', 'error_provenance': {'function': 'double', 'func_kwargs': {}, 'value': {'path': 'fail'}, 'value_path': 'path', 'error': 'Input must be an int'}, 'timestamp': ... 'traceback': '', 'provenance': {'function': 'double', 'func_kwargs': {}, 'value': {'path': 'fail'}, 'value_path': 'path'}}]
+    [{'output': 2, 'provenance': {'function': 'double', 'func_kwargs': {}, 'value': {'path': 1}, 'value_path': 'path'}}, {'error': 'Function double did not execute properly during map.', 'function': 'annotate_map_metadata', 'error_provenance': {'function': 'double', 'func_kwargs': {}, 'value': {'path': 'fail'}, 'value_path': 'path', 'error': 'Input must be an int'}, 'timestamp': ... 'traceback': ... 'provenance': {'function': 'double', 'func_kwargs': {}, 'value': {'path': 'fail'}, 'value_path': 'path'}}]
 
     Example of trying to call nonexistent function, `triple`
     >>> asyncio.run(handle_map({'double': double}, 'triple', [{'path': i} for i in range(2)], 'path'))
     Traceback (most recent call last):
       ...
-    learning_observer.communication_protocol.exception.DAGExecutionException: ('Could not find function `triple` in available functions.', 'handle_map', {'function_name': 'triple', 'available_functions': dict_keys(['double']), 'error': "'triple'"})
+    learning_observer.communication_protocol.exception.DAGExecutionException: ('Could not find function `triple` in available functions.', 'handle_map', {'function_name': 'triple', 'available_functions': dict_keys(['double']), 'error': "'triple'"}, ...)
     """
     if func_kwargs is None:
         func_kwargs = {}
@@ -424,7 +424,7 @@ async def handle_select(keys, fields):
                 value = get_nested_dict_value(resulting_value, f)
             except KeyError as e:
                 value = DAGExecutionException(
-                    f'KeyError: key not found',
+                    f'KeyError: key `{f}` not found in `{resulting_value.keys()}`',
                     inspect.currentframe().f_code.co_name,
                     {'target': resulting_value, 'key': f, 'exception': e}
                 ).to_dict()
@@ -481,7 +481,7 @@ def hack_handle_keys(function, STUDENTS=None, STUDENTS_path=None, RESOURCES=None
         fields = [
             {
                 learning_observer.stream_analytics.fields.KeyField.STUDENT: get_nested_dict_value(s, STUDENTS_path),  # TODO catch get_nested_dict_value errors
-                learning_observer.stream_analytics.helpers.EventField('doc_id'): get_nested_dict_value(r, RESOURCES_path)  # TODO catch get_nested_dict_value errors
+                learning_observer.stream_analytics.helpers.EventField('doc_id'): get_nested_dict_value(r, RESOURCES_path, '')  # TODO catch get_nested_dict_value errors
             } for s, r in zip(STUDENTS, RESOURCES)
         ]
         provenances = [
@@ -670,8 +670,9 @@ async def visit(node_name):
                 'error_path': error_path
             }
             error_texts = '\n'.join((f'  {e}' for e in _find_error_messages(error)))
+            tb = nodes[node_name]["error"].get("traceback", 'No traceback available')
             debug_log('ERROR:: Error occured within execution dag at '\
-                      f'{node_name}\n{nodes[node_name]["error"]["traceback"]}\n'\
+                      f'{node_name}\n{tb}\n'\
                       f'{error_texts}')
         else:
             nodes[node_name] = await dispatch_node(nodes[node_name])

learning_observer/learning_observer/stream_analytics/helpers.py

@@ -110,7 +110,7 @@ def make_key_from_json(js):
     ]
 
     if KeyField.STUDENT in js:
-        user_id = j[sKeyField.STUDENT]
+        user_id = js[KeyField.STUDENT]
 
     aggregator_functions = sum(
         [

learning_observer/learning_observer/util.py

@@ -151,7 +151,7 @@ def get_nested_dict_value(d, key_str=None, default=MissingType.Missing):
             d = d
         else:
             if default == MissingType.Missing:
-                raise KeyError(f'Key {key_str} not found in {d}')
+                raise KeyError(f'Key `{key_str}` not found in {d}')
             return default
     return d
 
@@ -165,11 +165,11 @@ def remove_nested_dict_value(d, key_str):
         if d is not None and key in d:
             d = d[key]
         else:
-            raise KeyError(f'Key {key_str} not found in {d}')
+            raise KeyError(f'Key `{key_str}` not found in {d}')
     if keys[-1] in d:
         return d.pop(keys[-1])
     else:
-        raise KeyError(f'Key {key_str} not found in {d}')
+        raise KeyError(f'Key `{key_str}` not found in {d}')
 
 
 def clean_json(json_object):

modules/wo_bulk_essay_analysis/wo_bulk_essay_analysis/assets/scripts.js

@@ -107,7 +107,7 @@ window.dash_clientside.bulk_essay_feedback = {
   /**
    * Sends data to server via websocket
    */
-  send_to_loconnection: async function (state, hash, clicks, query, systemPrompt, tags) {
+  send_to_loconnection: async function (state, hash, clicks, docSrc, docDate, docTime, query, systemPrompt, tags) {
     if (state === undefined) {
       return window.dash_clientside.no_update
     }
@@ -116,14 +116,16 @@ window.dash_clientside.bulk_essay_feedback = {
       const decoded = decode_string_dict(hash.slice(1))
       if (!decoded.course_id) { return window.dash_clientside.no_update }
 
-      decoded.gpt_prompt = ''
-      decoded.message_id = ''
+      decoded.gpt_prompt = '';
+      decoded.message_id = '';
+      decoded.doc_source = docSrc;
+      decoded.requested_timestamp = new Date(`${docDate}T${docTime}`).getTime().toString();
 
       const trig = window.dash_clientside.callback_context.triggered[0]
       if (trig.prop_id.includes('bulk-essay-analysis-submit-btn')) {
-        decoded.gpt_prompt = query
-        decoded.system_prompt = systemPrompt
-        decoded.tags = tags
+        decoded.gpt_prompt = query;
+        decoded.system_prompt = systemPrompt;
+        decoded.tags = tags;
       }
 
       const message = {

modules/wo_bulk_essay_analysis/wo_bulk_essay_analysis/dashboard/layout.py

@@ -4,6 +4,7 @@
 '''
 import dash_bootstrap_components as dbc
 from dash_renderjson import DashRenderjson
+import datetime
 import lo_dash_react_components as lodrc
 
 from dash import html, dcc, clientside_callback, ClientsideFunction, Output, Input, State, ALL
@@ -26,6 +27,10 @@
 
 advanced_collapse = f'{prefix}-advanced-collapse'
 system_input = f'{prefix}-system-prompt-input'
+# document source
+doc_src = f'{prefix}-doc-src'
+doc_src_date = f'{prefix}-doc-src-date'
+doc_src_timestamp = f'{prefix}-doc-src-timestamp'
 
 attachment_upload = f'{prefix}-attachment-upload'
 attachment_label = f'{prefix}-attachment-label'
@@ -64,7 +69,17 @@ def layout():
                 dbc.InputGroupText('System prompt:'),
                 dbc.Textarea(id=system_input, value=system_prompt)
             ]),
-            dcc.Store(id=attachment_store, data='')
+            html.Div([
+                dbc.Label('Document Source'),
+                dbc.RadioItems(options=[
+                    {'label': 'Latest Document', 'value': 'latest' },
+                    {'label': 'Specific Time', 'value': 'ts'},
+                ], value='latest', id=doc_src),
+                dbc.InputGroup([
+                    dcc.DatePickerSingle(id=doc_src_date, date=datetime.date.today()),
+                    dbc.Input(type='time', id=doc_src_timestamp, value=datetime.datetime.now().strftime("%H:%M"))
+                ])
+            ])
         ], label='Advanced', id=advanced_collapse, is_open=False),
     ])
 
@@ -87,7 +102,8 @@ def layout():
         dbc.CardFooter([
             html.Small(id=attachment_warning_message, className='text-danger'),
             dbc.Button('Save', id=attachment_save, color='primary', n_clicks=0, class_name='float-end')
-        ])
+        ]),
+        dcc.Store(id=attachment_store, data='')
     ], class_name='h-100')
 
     # query creator panel
@@ -136,16 +152,27 @@ def layout():
     return dcc.Loading(cont)
 
 
+# disbale document date/time options
+clientside_callback(
+    ClientsideFunction(namespace='clientside', function_name='disable_doc_src_datetime'),
+    Output(doc_src_date, 'disabled'),
+    Output(doc_src_timestamp, 'disabled'),
+    Input(doc_src, 'value')
+)
+
 # send request on websocket
 clientside_callback(
     ClientsideFunction(namespace='bulk_essay_feedback', function_name='send_to_loconnection'),
     Output(websocket, 'send'),
     Input(websocket, 'state'),  # used for initial setup
     Input('_pages_location', 'hash'),
     Input(submit, 'n_clicks'),
+    Input(doc_src, 'value'),
+    Input(doc_src_date, 'date'),
+    Input(doc_src_timestamp, 'value'),
     State(query_input, 'value'),
     State(system_input, 'value'),
-    State(tag_store, 'data')
+    State(tag_store, 'data'),
 )
 
 # enable/disabled submit based on query

modules/wo_common_student_errors/wo_common_student_errors/assets/scripts.js

@@ -24,7 +24,7 @@ const categoryColors = {
 }
 
 window.dash_clientside.common_student_errors = {
-  send_to_loconnection: function (state, hash) {
+  send_to_loconnection: function (state, hash, docSrc, docDate, docTime) {
     /**
      * When the hash of the URL changes, we send an updated query
      * to the Learning Observer server.
@@ -46,6 +46,8 @@ window.dash_clientside.common_student_errors = {
       } else {
         decoded.student_id = []
       }
+      decoded.doc_source = docSrc;
+      decoded.requested_timestamp = new Date(`${docDate}T${docTime}`).getTime().toString();
       const message = {
         wo: {
           execution_dag: 'writing_observer',

modules/wo_common_student_errors/wo_common_student_errors/dashboard/layout.py

@@ -6,6 +6,7 @@
 # package imports
 import dash_bootstrap_components as dbc
 from dash_renderjson import DashRenderjson
+import datetime
 import lo_dash_react_components as lodrc
 import plotly.express as px
 import writing_observer.languagetool
@@ -27,6 +28,11 @@
 alert_text = f'{prefix}-alert-text'
 alert_error_dump = f'{prefix}-alert-error-dump'
 
+# document source
+doc_src = f'{prefix}-doc-src'
+doc_src_date = f'{prefix}-doc-src-date'
+doc_src_timestamp = f'{prefix}-doc-src-timestamp'
+
 # error per text length items
 error_per_length = f'{prefix}-errors-per-length-graph'
 error_per_length_tooltip = f'{prefix}-errors-per-length-tooltip'
@@ -49,6 +55,17 @@ def layout():
 
     tooltip = dcc.Tooltip(id=error_per_length_tooltip, direction='bottom')
     overall_view = html.Div([
+        html.Div([
+            dbc.Label('Document Source'),
+            dbc.RadioItems(options=[
+                {'label': 'Latest Document', 'value': 'latest' },
+                {'label': 'Specific Time', 'value': 'ts'},
+            ], value='latest', id=doc_src),
+            dbc.InputGroup([
+                dcc.DatePickerSingle(id=doc_src_date, date=datetime.date.today()),
+                dbc.Input(type='time', id=doc_src_timestamp, value=datetime.datetime.now().strftime("%H:%M"))
+            ])
+        ]),
         activity.layout,
         dcc.Graph(
             id=error_per_length,
@@ -99,13 +116,25 @@ def layout():
     return dcc.Loading(cont)
 
 
+
+# disbale document date/time options
+clientside_callback(
+    ClientsideFunction(namespace='clientside', function_name='disable_doc_src_datetime'),
+    Output(doc_src_date, 'disabled'),
+    Output(doc_src_timestamp, 'disabled'),
+    Input(doc_src, 'value')
+)
+
 # send request to LOConnection
 clientside_callback(
     ClientsideFunction(namespace='common_student_errors', function_name='send_to_loconnection'),
     Output(websocket, 'send'),
     Output(individual.prefix, 'className'),
     Input(websocket, 'state'),  # used for initial setup
-    Input('_pages_location', 'hash')
+    Input('_pages_location', 'hash'),
+    Input(doc_src, 'value'),
+    Input(doc_src_date, 'date'),
+    Input(doc_src_timestamp, 'value'),
 )
 
 # Update the url's hash based on errors per text length graph's selectedData

modules/wo_document_list/wo_document_list/assets/scripts.js

@@ -107,7 +107,8 @@ window.dash_clientside.document_list = {
       if (!decoded.course_id) { return window.dash_clientside.no_update }
 
       decoded.assignment_id = assignment || ''
-      decoded.tag_path = tag ? `tags.${tag}` : ''
+
+      decoded.tag_path = tag ? `tags.${tag}` : 'tags'
 
       const message = {
         wo: {

modules/wo_highlight_dashboard/wo_highlight_dashboard/assets/scripts.js

@@ -59,6 +59,13 @@ window.dash_clientside.clientside = {
     return [text, true, data];
   },
 
+  disable_doc_src_datetime: function (value) {
+    if (value === 'ts') {
+      return [false, false];
+    }
+    return [true, true];
+  },
+
     change_sort_direction_icon: function(sort_check, sort_values) {
         // updates UI elements, does not handle sorting
         // based on the current sort, set the sort direction icon and sort text
@@ -401,7 +408,7 @@ window.dash_clientside.clientside = {
         return true;
     },
 
-    send_options_to_server: function(types, metrics, highlights, indicators, sort_by, course_id) {
+    send_options_to_server: function(types, metrics, highlights, indicators, sort_by, course_id, doc_src, doc_date, doc_time) {
         // Send selected options to the server 
         // TODO work on protocol for communicating with the 
         //
@@ -410,15 +417,21 @@ window.dash_clientside.clientside = {
         // Input(settings.metric_checklist, 'value'),
         // Input(settings.highlight_checklist, 'value'),
         // Input(settings.indicator_checklist, 'value')
-        // Input(settings.sort_by_checklist, 'value')
+        // Input(settings.sort_by_checklist, 'value'),
+        // Input(course_store, 'data'),
+        // Input(settings.doc_src, 'value'),
+        // Input(settings.doc_src_date, 'date'),
+        // Input(settings.doc_src_timestamp, 'value')
         const options = metrics.concat(highlights).concat(indicators).concat(sort_by);
         const message = {
             docs_with_nlp: {
                 execution_dag: 'writing_observer',
                 target_exports: ['docs_with_nlp_annotations'],
                 kwargs: {
                     course_id: course_id,
-                    nlp_options: options
+                    nlp_options: options,
+                    doc_source: doc_src,
+                    requested_timestamp: new Date(`${doc_date}T${doc_time}`).getTime().toString()
                 }
             }
         }