updated gpsr break down

Author: Oscar-gg

docs/development/integration/Task Breakdown/gpsr/commands.md

@@ -1,4 +1,4 @@
-# Command Brake Down
+# Command Break Down
 
 Go to the dishwasher then find an apple and get it and bring it to the waving person in the office
 
@@ -10,6 +10,9 @@ Go to the dishwasher then find an apple and get it and bring it to the waving pe
 -> pick(apple)
 -> go(office)
 -> find(waving_person)
+----> gaze(n predefined points)
+----> detect(waving_person)
+----> approach_point(waving_person)
 -> give(apple)
 
 Locate a waving person in the kitchen and answer a question
@@ -21,42 +24,43 @@ Locate a waving person in the kitchen and answer a question
 ----> approach_point(waving_person)
 -> say(ask_question)
 -> hear(user)
--> say(answer)
+-> contextual_say(answer)
 
 Meet Axel in the bedroom and answer a quiz
 -> go(bedroom)
 -> find(Axel)
 ----> gaze(n predefined points)
-----> detect(Axel)
-----> approach_point(Axel)
+----> detect(person)
+----> if person[i] == Axel then approach_point(Axel)
+----> else: for person in persons: approach_point(person) -> ask if is Axel
+----> save_to_context(Axel)
 -> say(ask_question)
 -> hear(user)
--> say(answer)
+-> contextual_say(answer)
 
 Tell me how many waving persons are in the kitchen
 -> go(kitchen)
 -> count(waving_person)
 ----> gaze(n predefined points)
 ----> detect(waving person)
 -> go(task_origin)
--> say(answer)
+-> contextual_say(answer)
 
 Tell me the gesture of the person at the tv stand
 -> go(tv_stand)
 -> find(person)
 ----> gaze(n predefined points)
 ----> detect(person)
 -> go(task_origin)
--> say(gesture)
+-> contextual_say(gesture)
 
 Say your teams affiliation to the waving person in the bathroom
 -> go(bathroom)
 -> find(waving_person)
 ----> gaze(n predefined points)
 ----> detect(waving person)
 ----> approach_point(waving_person)
--> fetch(affiliation)
--> say(affiliation)
+-> contextual_say(affiliation)
 
 Answer the quiz of the person raising their right arm in the living room
 -> go(living_room)
@@ -66,15 +70,18 @@ Answer the quiz of the person raising their right arm in the living room
 ----> approach_point(person_raising_right_arm)
 -> say(ask_quiz)
 -> hear(user)
--> say(answer)
+-> contextual_say(answer)
 
 Follow Jane from the desk to the kitchen
 -> go(desk)
 -> find(Jane)
 ----> gaze(n predefined points)
 ----> detect(person)
-----> approach_point(person)
-----> confirm(person)
+----> if person[i] == Jane then approach_point(Jane)
+----> else: for person in persons: approach_point(person) -> ask if is Jane
+----> save_to_context(Jane)
+
+-> say(intent)
 -> follow_person_until(Jane, kitchen)
 
 Escort Axel from the trashbin to the pantry
@@ -83,8 +90,11 @@ Escort Axel from the trashbin to the pantry
 -> find(Axel)
 ----> gaze(n predefined points)
 ----> detect(Axel)
-----> approach_point(Axel)
-----> confirm(person)
+----> if person[i] == Axel then approach_point(Axel)
+----> else: for person in persons: approach_point(person) -> ask if is Axel
+----> save_to_context(Axel)
+
+-> say(intent)
 -> escort_to(pantry)
 
 Take the person pointing to the right from the armchair to the exit
@@ -93,6 +103,7 @@ Take the person pointing to the right from the armchair to the exit
 ----> gaze(n predefined points)
 ----> detect(person_pointing_right)
 ----> approach_point(person_pointing_right)
+-> say(intent)
 -> escort_to(exit)
 
 Take the person wearing a yellow t shirt from the side tables to the office
@@ -101,6 +112,7 @@ Take the person wearing a yellow t shirt from the side tables to the office
 ----> gaze(n predefined points)
 ----> detect(person_yellow_tshirt)
 ----> approach_point(person_yellow_tshirt)
+-> say(intent)
 -> escort_to(office)
 
 Say hello to the person wearing a gray t shirt in the living room and answer a quiz
@@ -112,66 +124,80 @@ Say hello to the person wearing a gray t shirt in the living room and answer a q
 -> say(hello)
 -> say(ask_quiz)
 -> hear(user)
--> say(answer)
+-> contextual_say(answer)
 
 Say hello to Morgan in the kitchen and answer a question
 -> go(kitchen)
 -> find(Morgan)
 ----> gaze(n predefined points)
 ----> detect(Morgan)
-----> approach_point(Morgan)
+----> if person[i] == Morgan then approach_point(Morgan)
+----> else: for person in persons: approach_point(Morgan) -> ask if is Morgan
+----> save_to_context(Morgan)
+
 -> say(hello)
 -> say(ask_question)
 -> hear(user)
--> say(answer)
+-> contextual_say(answer)
 
+Note: in the second find, the person should be already known i/e found person[i] == Jane
 Meet Jane at the entrance then locate them in the kitchen
 -> go(entrance)
 -> find(Jane)
 ----> gaze(n predefined points)
 ----> detect(Jane)
-----> approach_point(Jane)
+----> if person[i] == Jane then approach_point(Jane)
+----> else: for person in persons: approach_point(Jane) -> ask if is Jane
+----> save_to_context(Jane)
 -> go(kitchen)
 -> find(Jane)
 ----> gaze(n predefined points)
 ----> detect(Jane)
+----> if person[i] == Jane then approach_point(Jane)
+----> else: for person in persons: approach_point(Jane) -> ask if is Jane
+----> save_to_context(Jane)
+-> say(intent finished)
 
 Tell me how many people in the office are wearing yellow coats
 -> go(office)
 -> count(person_yellow_coat)
 ----> gaze(n predefined points)
 ----> detect(person_yellow_coat)
+----> save_to_context(count)
+
 -> go(task_origin)
--> say(answer)
+-> contextual_say(answer)
 
 Tell me how many people in the living room are wearing gray coats
 -> go(living_room)
 -> count(person_gray_coat)
 ----> gaze(n predefined points)
 ----> detect(person_gray_coat)
+----> save_to_context(count)
 -> go(task_origin)
--> say(answer)
+-> contextual_say(answer)
 
 Tell the gesture of the person at the side tables to the person at the tv stand
 -> go(side_tables)
 -> find(person)
 ----> gaze(n predefined points)
 ----> detect(person)
--> fetch(gesture)
+----> save_to_context(gesture)
+
 -> go(tv_stand)
 -> find(person)
 ----> gaze(n predefined points)
 ----> detect(person)
--> say(gesture)
+-> contextual_say(gesture)
 
 Follow the standing person in the living room
 -> go(living_room)
 -> find(standing_person)
 ----> gaze(n predefined points)
 ----> detect(standing_person)
 ----> approach_point(standing_person)
-----> confirm(person)
--> follow_person_until(standing_person, destination)
+-> say(intent)
+-> follow_person_until_canceled(standing_person)
 
 Go to the kitchen then locate a snack and get it and deliver it to Adel in the bathroom
 -> go(kitchen)
@@ -184,7 +210,10 @@ Go to the kitchen then locate a snack and get it and deliver it to Adel in the b
 -> find(Adel)
 ----> gaze(n predefined points)
 ----> detect(Adel)
-----> approach_point(Adel)
+----> if person[i] == Adel then approach_point(Adel)
+----> else: for person in persons: approach_point(Adel) -> ask if is Adel
+----> save_to_context(Adel)
+-> say(intent)
 -> give(snack)
 
 Take a cleaning supply from the kitchen table and bring it to the standing person in the kitchen
@@ -198,6 +227,7 @@ Take a cleaning supply from the kitchen table and bring it to the standing perso
 ----> gaze(n predefined points)
 ----> detect(standing_person)
 ----> approach_point(standing_person)
+-> say(intent)
 -> give(cleaning_supply)
 
 Locate a cleaning supply in the bedroom then take it and bring it to the standing person in the bedroom
@@ -211,24 +241,27 @@ Locate a cleaning supply in the bedroom then take it and bring it to the standin
 ----> gaze(n predefined points)
 ----> detect(standing_person)
 ----> approach_point(standing_person)
+-> say(intent)
 -> give(cleaning_supply)
 
 Tell me how many fruits there are on the bed
 -> go(bed)
 -> count(fruits)
 ----> gaze(n predefined points)
 ----> detect(fruits)
+----> save_to_context(count)
 -> go(task_origin)
--> say(answer)
+-> contextual_say(answer)
 
 TODO: how to detect biggest object?
 Tell me what is the biggest object on the kitchen table
 -> go(kitchen_table)
 -> find(biggest_object)
 ----> gaze(n predefined points)
 ----> detect(biggest_object)
+----> save_to_context(objects)
 -> go(task_origin)
--> say(biggest_object)
+-> contextual_say(biggest_object)
 
 Bring me a soccer ball from the dishwasher
 -> go(dishwasher)
@@ -246,9 +279,10 @@ Tell me what is the lightest toy on the bookshelf
 -> find(lightest_toy)
 ----> gaze(n predefined points)
 ----> detect(toys)
-----> determine_lightest(toys)
+----> save_to_context(toys)
+
 -> go(task_origin)
--> say(lightest_toy)
+-> contextual_say(lightest_toy)
 
 ## Todo ideas
 
@@ -257,7 +291,7 @@ TODO: how to detect lightest object that is a toy?
 - When in the find method, insert all objects matching description to a db.
 - In the db, store previously know object properties.
 - When determining the lightest object, compare the objects in the db and return the lightest one.
-  - Perhaps, implement a contextual_say method that fetches context to answer questies?
+  - Perhaps, implement a contextual_say method that fetches context to answer queries?
   - Let the contextual say choose between the databases
 
 TODO: how to detect biggest object?
@@ -266,12 +300,12 @@ TODO: how to detect biggest object?
 
 TODO: how to handle finds with different requirements?
 
-- i.e. find by name, by posture, by object recognizable with model
+- i.e. find by name, by posture, by object recognizable with model?
 
 TODO: deux ex machina integration
 
 -
 
 TODO: confirmation implementation
 
-- How to add confirmation and fallbacks to the subcommands?
+- How to add confirmation and fallbacks to the subcommands? add confirmation within the subcommand or before/after the subcommands?