Chatbot_ner designed to extract essential entities or information like date, time, numbers, texts (city, organisation name) from textual data. These APIs are built keeping in mind to plug it in for conversational AI applications.
Following are the different entity types we currently support.
| Entity Type | Description | Examples |
|---|---|---|
| Time | Detects time in various formats from given text. | tomorrow morning at 5, कल सुबह ५ बजे, kal subah 5 baje |
| Date | Detects date in various formats from given text. | 28-12-2096, 09th Nov 2014, Tomorrow, next monday, agle somvar, अगले सोमवार |
| Number | Detects number from the text. | 50 rs per person, ५ किलो चावल, मुझे एक लीटर ऑइल चाहिए |
| Phone number | Detects phone numbers in given text. | +919222222222 |
| Detects email addresses in given text. | [email protected] | |
| Text | Detect custom entities in text string using full text search in Datastore or based on contextual model | pizza, मुंबई |
| PNR-Number | Detect PNR (serial) codes in given text. | My flight PNR is 4SGX3E |
| Regex | Detect entities using custom regex patterns | My flight PNR is 4SGX3E Please apply CASH20 coupon code |
Following are the list of parameters accepted by APIs used for detection:
-
message: Message on which detection logic needs to run. It is an unstructured text from which entity needs to be extracted. For example, "I want to order pizza".
-
entity_name: Name of the entity. This parameter primarily defines the key of the output dictionary.
For example, in Text detection logic the entity name is based on dictionary name of entities we need to detect.
- For detecting:
- cuisine entity_name will be "cuisine"
- dish entity_name will be "dish"
For other entities like
dateortimeentity_name, detection logic works solely on their entity_types. - For detecting:
-
structured_value (optional): It is a value which is obtained from the structured text. For example, UI elements like form, payload, etc. See below image for reference used for detecting test name from form key
Test name
-
fallback_value (optional): It is a fallback value. If the detection logic fails to detect any value either from structured_value or message then we return a fallback_value passed in request as an output.
For example, if user says "Nearby ATMs". In this example user has not provided any information about his location in the chat but, we can pass a fallback_value that will contain its location that can be obtained from its profile or third party apis (like geopy, etc).
-
source_language: It is the language in which user message is passed. We have to pass ISO 639-1 code of the language here.
-
bot_message (optional): previous message from a bot/agent. This is an important parameter, many times the entity value relies on the message from the bot/agent i.e. what is bot saying? or asking?
For example, bot might ask for departure date to book a flight and user might reply with a date. Now, it is difficult to disambiguate whether its departure date or arrival date unless, we know what bot is asking for?
bot: Please help me with date of departure? user: 23rd March
The output of entities detected will be stored in a list of dictionary containing the following structure:
[
{
"entity_value": entity_value,
"detection": detection_method,
"original_text": original_text
},
]
Consider the following example for detailed explanation:
"I want to order two litre pepsi."
- entity_value: This will store the value of entity (i.e entity value) that is detected. The output will be in dictionary format. For example,
{"value": "2", "unit": "quantity"}. - detection: This indicates how the entity was detected. Possibles values are
- message : If entity get detected from message key.
- structured value: If entity get detected from structured value.
- fallback value: If entity get detected from fallback value.
- original_text: This store the part of the text that gets detected as an entity. For example,
two.
Example 1:
input:
message = u'I want to order 2 burgers from mainland china at 3 pm '
entity_name = 'restaurant'
structured_value = None
structured_value_verification = 0
fallback_value = None
bot_message = None
output:
[
{
"detection": "message",
"original_text": "2",
"entity_value": { "value": "2", "unit": null},
"language": "en"
},
{
"detection": "message",
"original_text": "3 pm",
"entity_value": { "mm": 0, "hh": 3, "nn": "pm"},
"language": "en"
}
]
Example 2:
// If today is 19th feb
input:
message = u'I have my maths exam next Saturday.'
entity_name = 'date'
structured_value = None
structured_value_verification = 0
fallback_value = None
bot_message = None
output:[
{
"detection": "message",
"original_text": "inferno",
"entity_value": {"value": {"mm":03, "yy": 2019, "dd": 02, "type": "date"}},
"language": "en"
},
]Please, have a look at installation steps guide to install Chatbot NER on your system.
After following above steps, you can test the APIs either using django shell or by making curl get request.
To get into django shell you need to execute the below commands.
- To get inside container, first run
docker exec -it <container_name> bash - Then, inside main project directory run
python manage.py shell
Below are the APIs example for each of the entity types mentioned above.
The Time detector module has the capability to detect time from text in multiple languages. It can detect time in 12/24 hr format and also detect text with time difference specified. for example - wake me up after 10 mins.
Currently time detection support has been provided in different languages - English, Hindi, Marathi, Bengali, Gujrati, Tamil. It also supports latin script of given languages.
To add support for new languages or to add custom patterns in particular language please go through the Time Detector readme here
Note - This module has been updated to v2 version of chatbot_ner.
API Example:
-
Example 1: Detect time from English text which contains time mentions in 24 hours format, 12 hours format and relative time format
-
Django Shell
message = u"John arrived at the bus stop at 13:50 hrs, expecting the bus to be there in 15 mins. But the bus was scheduled for 12:30 pm" entity_name = 'time' structured_value = None fallback_value = None bot_message = None timezone = 'UTC' source_language='en' from ner_v2.detectors.temporal.time.time_detection import TimeDetector detector = TimeDetector(entity_name=entity_name, language=source_language, timezone=timezone) output = detector.detect(message=message, entity_name=entity_name, structured_value=structured_value, fallback_value=fallback_value) print(output)
-
CURL command
curl -G -i "http://localhost:8081/v2/time/?&entity_name=time&timezone=UTC&source_language=en" \ --data-urlencode "structured_value=" \ --data-urlencode "fallback_value=" \ --data-urlencode "bot_message=" \ --data-urlencode "message=John arrived at the bus stop at 13:50 hrs, expecting the bus to be there in 15 mins. But the bus was scheduled for 12:30 pm"
Output:
{ "data": [ { "detection": "message", "original_text": "12:30 pm", "entity_value": { "mm": 30, "hh": 12, "nn": "pm" }, "language": "en" }, { "detection": "message", "original_text": "in 15 mins", "entity_value": { "mm": 15, "hh": 0, "nn": "df" }, "language": "en" }, { "detection": "message", "original_text": "13:50", "entity_value": { "mm": 50, "hh": 13, "nn": "hrs" }, "language": "en" } ] }
-
-
Example 2: Detect time[Hindi] from text containing 24 hrs, 12 hrs and time difference text format
-
Django Shell
message = u"राजू का बस १३:५० को बस स्टॉप से निकला और १५ मिनट में यहाँ पहुंच जाएगा और गोवा को शाम में बारह बजकर ३० मिनट पैर पहुंचेगा" entity_name = 'time' structured_value = None fallback_value = None bot_message = None timezone = 'UTC' source_language='hi' from ner_v2.detectors.temporal.time.time_detection import TimeDetector detector = TimeDetector(entity_name=entity_name, language=source_language, timezone=timezone) output = detector.detect(message=message, entity_name=entity_name, structured_value=structured_value, fallback_value=fallback_value) print(output)
-
CURL command
curl -G -i "http://localhost:8081/v2/time/?&entity_name=time&timezone=UTC&source_language=hi" \ --data-urlencode "structured_value=" \ --data-urlencode "fallback_value=" \ --data-urlencode "bot_message=" \ --data-urlencode "message=राजू का बस १३:५० को बस स्टॉप से निकला और १५ मिनट में यहाँ पहुंच जाएगा और गोवा को शाम में बारह बजकर ३० मिनट पैर पहुंचेगा"
Output:
{ "data": [ { "detection": "message", "original_text": "१३:५०", "entity_value": { "mm": 50, "hh": 13, "nn": "hrs", "time_type": null }, "language": "hi" }, { "detection": "message", "original_text": "१५ मिनट में", "entity_value": { "mm": 15, "hh": 0, "nn": "df" }, "language": "hi" }, { "detection": "message", "original_text": "बारह बजकर ३० मिनट", "entity_value": { "mm": 30, "hh": 12, "nn": "hrs" }, "language": "hi" } ] }
-
The Date detector module has the capability to detect various form of dates from text in multiple languages. It can detect date from following patterns:
- Day month year format - 12 feb 2018, 2nd Jan 2019, 12/11/2019, 12-jan-2019
- Day month - 12 feb, 12/12
- Weekday reference - Comming monday, next sunday
- Reference day month - 2nd of next month, 2nd sunday of coming month
- Current day reference - tomorrow, yesterday, day after tomorrow
Currently date detection support has been provided in different languages - English, Hindi, Marathi, Bengali, Gujrati, Tamil. It also supports latin script of given languages.
To add support for new languages or to add custom patterns in particular language please go through the Date Detector readme here
Note - This module has been updated to v2 version of chatbot_ner
API Examples:
-
Example 1: Detecting day month format date [English] from user message
Use the timezone parameter to pass your current timezone to date detection
-
Django Shell:
message = u"set me reminder on 23rd december" entity_name = 'date' structured_value = None fallback_value = None bot_message = None timezone='UTC' source_language='en' past_date_referenced=False # flag to check if the date reference lies in past or future. For Example - In hindi "Kal"corresonds to both tomorrow and yesterday. So this flag will determines actual referenced date. from ner_v2.detectors.temporal.date.date_detection import DateAdvanceDetector detector = DateAdvanceDetector(entity_name=entity_name, language=source_language, timezone=timezone, past_date_referenced=past_date_referenced) output = detector.detect(message=message, entity_name=entity_name, structured_value=structured_value, fallback_value=fallback_value) print(output)
-
CURL:
curl -G -i "http://localhost:8081/v2/date/?&entity_name=date&timezone=UTC&source_language=en" \ --data-urlencode "structured_value=" \ --data-urlencode "fallback_value=" \ --data-urlencode "bot_message=" \ --data-urlencode "message=set me reminder on 23rd december"
Output:
{ "data": [ { "detection": "message", "original_text": "23rd december", "entity_value": { "end_range": false, "from": false, "normal": true, "value": { "mm": 12, "yy": 2019, "dd": 23, "type": "date" }, "to": false, "start_range": false } } ] }
-
-
Example 2: Detecting referenced date [Hindi] from user message
Use the timezone parameter to pass your current timezone to date detection
-
Django Shell:
message = u"मुझे कल सुबह ५ बजे उठा देना" entity_name = 'date' structured_value = None fallback_value = None bot_message = None timezone='UTC' source_language='hi' past_date_referenced=False from ner_v2.detectors.temporal.date.date_detection import DateAdvanceDetector detector = DateAdvanceDetector(entity_name=entity_name,language=source_language, timezone=timezone, past_date_referenced=past_date_referenced) output = detector.detect(message=message, entity_name=entity_name, structured_value=structured_value, fallback_value=fallback_value) print(output)
-
CURL:
curl -G -i "http://localhost:8081/v2/date/?&entity_name=date&timezone=UTC&past_date_referenced=false&source_language=hi" \ --data-urlencode "structured_value=" \ --data-urlencode "fallback_value=" \ --data-urlencode "bot_message=" \ --data-urlencode "message=मुझे कल सुबह ५ बजे उठा देना"
Output:
/* Assuming today's date is 27 June 2019*/ { "data": [ { "detection": "message", "original_text": "कल", "entity_value": { "end_range": false, "from": false, "normal": true, "value": { "mm": 6, "yy": 2019, "dd": 28, "type": "date" }, "to": false, "start_range": false }, "language": "hi" } ] }
-
-
Example 3: Detecting referenced weekday [Hindi] from user message
Use the timezone parameter to pass your current timezone to date detection
-
Django Shell:
message = u"आने वाले सोमवार को मेरा मैथ्स का एग्जाम है" entity_name = 'date' structured_value = None fallback_value = None bot_message = None timezone='UTC' source_language='hi' past_date_referenced=False from ner_v2.detectors.temporal.date.date_detection import DateAdvanceDetector detector = DateAdvanceDetector(entity_name=entity_name,language=source_language, timezone=timezone, past_date_referenced=past_date_referenced) output = detector.detect(message=message, entity_name=entity_name, structured_value=structured_value, fallback_value=fallback_value) print(output)
-
CURL:
curl -G -i "http://localhost:8081/v2/date/?&entity_name=date&timezone=UTC&past_date_referenced=false&source_language=hi" \ --data-urlencode "structured_value=" \ --data-urlencode "fallback_value=" \ --data-urlencode "bot_message=" \ --data-urlencode "message=आने वाले सोमवार को मेरा मैथ्स का एग्जाम है"
Output:
/* Assuming today's date is 27 June 2019*/ { "data": [ { "detection": "message", "original_text": "सोमवार", "entity_value": { "end_range": false, "from": false, "normal": true, "value": { "mm": 7, "yy": 2019, "dd": 1, "type": "date" }, "to": false, "start_range": false }, "language": "hi" } ] }
-
The Number detector module has the capability to detect number or number word from text in multiple languages. The detector supports an additional feature of detecting units along with number if given in text. For example - for a given text 5 kg, this module will return 5 as detected value and kg as detected unit. It also has the capability to detect only certain type of numbers like currency or temperature type numbers by specifying the unit type.
You can find the current supported unit_types in english here
Currently number detection support has been provided for 6 different languages - English, Hindi, Marathi, Bengali, Gujrati, Tamil. It also supports latin script of given languages.
To add support for new languages or to add custom patterns in particular language please go through the Number Detector readme here
Note - This module has been updated to v2 version of chatbot_ner.
API Example:
-
Example 1: Detecting number[English] without unit in message
-
Django Shell
# For a sample query with following parameters message=u"i want to purchase 30 units of mobile abd 40 units of television" entity_name='number' structured_value=None fallback_value=None bot_message=None min_number_digits=1 # minimum number of digit max_number_digits=6 # maximum number of digit source_language='en' # here language will be ISO 639-1 code unit_type=None # this restrict the number detector to detect particular number type only. from ner_v2.detector.number.number.number_detection import NumberDetector detector = NumberDetector(entity_name=entity_name, language=source_language, unit_type=None) detector.set_min_max_digits(min_digit=min_number_digits, max_digit=max_number_digits) output = detector.detect(message=message,structured_value=structured_value, fallback_value=fallback_value, bot_message=bot_message) print(output)
-
CURL command:
curl -G -i "http://localhost:8081/v2/number/?&entity_name=number_of_unit&min_number_digits=1&max_number_digits=6&unit_type=&source_language=en" \ --data-urlencode "structured_value=" \ --data-urlencode "fallback_value=" \ --data-urlencode "bot_message=" \ --data-urlencode "message=i want to purchase 30 units of mobile abd 40 units of television"
Output:
{ "data": [ { "detection": "message", "original_text": "30", "entity_value": { "unit": null, "value": "30" }, "language": "en" }, { "detection": "message", "original_text": "40", "entity_value": { "unit": null, "value": "40" }, "language": "en" } ] }
-
-
Example 2: Detecting number[Hindi] without unit in message
-
Django Shell
# For a sample query with following parameters message=u"मुझे ३० रूपए आटा का और ३ हजार का चीनी देना" entity_name='number' structured_value=None fallback_value=None bot_message=None min_number_digits=1 # minimum number of digit max_number_digits=6 # maximum number of digit source_language='hi' # here language will be ISO 639-1 code unit_type="weight" # this restrict the number detector to detect only weight type number entity from ner_v2.detector.number.number.number_detection import NumberDetector detector = NumberDetector(entity_name=entity_name, language=source_language, unit_type=None) detector.set_min_max_digits(min_digit=min_number_digits, max_digit=max_number_digits) output = detector.detect(message=message,structured_value=structured_value, fallback_value=fallback_value, bot_message=bot_message) print(output)
-
CURL command:
curl -G -i "http://localhost:8081/v2/number/?&entity_name=number_of_unit&min_number_digits=1&max_number_digits=6&unit_type=&source_language=hi" \ --data-urlencode "structured_value=" \ --data-urlencode "fallback_value=" \ --data-urlencode "bot_message=" \ --data-urlencode "message=मुझे ३० रूपए आटा का और ३ हजार का चीनी देना"
Output:
{ "data": [ { "detection": "message", "original_text": "३०", "entity_value": { "unit": null, "value": "30" }, "language": "hi" }, { "detection": "message", "original_text": "३ हजार", "entity_value": { "unit": null, "value": "3000" }, "language": "hi" } ] }
-
-
Example 3: Detecting number[Hindi in latin script] without unit in message
-
Django Shell
# For a sample query with following parameters message=u"mujhe 30 kilo aata aur 2 hajaar ka chini dena aur teen sau ka chawal" entity_name='number' structured_value=None fallback_value=None bot_message=None min_number_digits=1 # minimum number of digit max_number_digits=6 # maximum number of digit source_language='hi' # here language will be ISO 639-1 code unit_type=None # this restrict the number detector to detect particular number type only. from ner_v2.detector.number.number.number_detection import NumberDetector detector = NumberDetector(entity_name=entity_name, language=source_language, unit_type=None) detector.set_min_max_digits(min_digit=min_number_digits, max_digit=max_number_digits) output = detector.detect(message=message,structured_value=structured_value, fallback_value=fallback_value, bot_message=bot_message) print(output)
-
CURL command:
curl -G -i "http://localhost:8081/v2/number/?&entity_name=number_of_unit&min_number_digits=1&max_number_digits=6&unit_type=&source_language=hi" \ --data-urlencode "structured_value=" \ --data-urlencode "fallback_value=" \ --data-urlencode "bot_message=" \ --data-urlencode "message=mujhe 30 kilo aata aur 2 hajaar ka chini dena aur teen sau ka chawal"
Output:
{ "data": [ { "detection": "message", "original_text": "30", "entity_value": { "unit": null, "value": "30" }, "language": "hi" }, { "detection": "message", "original_text": "2 hajaar", "entity_value": { "unit": null, "value": "2000" }, "language": "hi" }, { "detection": "message", "original_text": "teen sau", "entity_value": { "unit": null, "value": "300" }, "language": "hi" } ] }
-
-
Example 4: Detecting number[English] with unit in message
-
Django Shell
# For a sample query with following parameters message=u"i want more than Rupees 20k and 10 pendrive" entity_name='number' structured_value=None fallback_value=None bot_message=None min_number_digits=1 max_number_digits=6 source_language='en' # here language will be ISO 639-1 code unit_type='currency' # this restrict the number detector to detect particular number type only. from ner_v2.detector.number.number.number_detection import NumberDetector detector = NumberDetector(entity_name=entity_name, language=source_language, unit_type=None) detector.set_min_max_digits(min_digit=min_number_digits, max_digit=max_number_digits) output = detector.detect(message=message,structured_value=structured_value, fallback_value=fallback_value, bot_message=bot_message) print(output)
-
CURL command:
curl -G -i "http://localhost:8081/v2/number/?&entity_name=number_of_unit&min_number_digits=1&max_number_digits=6&unit_type=currency&source_language=en" \ --data-urlencode "structured_value=" \ --data-urlencode "fallback_value=" \ --data-urlencode "bot_message=" \ --data-urlencode "message=i want more than Rupees 20k and 10 pendrive"
Output:
{ "data": [ { "detection": "message", "original_text": "rupees 20k", "entity_value": { "unit": "rupees", "value": "20000" }, "language": "en" } ] } /* here 40 is not detected as the unit_type specified is currency, Hence it will only detect numbers with currencies mentioned as unit */
The Phone Number Detector has the capability to detect phone numbers from within the given text. The detector has the ability to handle multilanguage text. Additionally, this detector is scaled to handle domestic as well as international phone numbers.
This module has been updated to v2 version of chatbot_ner and is language agnostic.
API Examples:
-
Example 1: Detecting phone number from message
-
Django Shell:
message = u'send a message on 91 9820334455' entity_name = 'phone_number' structured_value = None fallback_value = None bot_message = None source_langauge='en' # here language will be ISO 639-1 code from ner_v2.detectors.pattern.phone_number.phone_number_detection import PhoneDetector detector = PhoneDetector(language=source_langauge, entity_name=entity_name) output = detector.detect(message=message, entity_name=entity_name, structured_value=structured_value, fallback_value=fallback_value, bot_message=bot_message,language=source_language) print(output)
-
CURL command:
curl -G -i "http://localhost:8081/v2/phone_number/?&entity_name=phone_number&source_language=en" \ --data-urlencode "structured_value=" \ --data-urlencode "fallback_value=" \ --data-urlencode "bot_message=" \ --data-urlencode "message=send a message on 91 9820334455"
**Output: **
{ "data": [ { "detection": "message", "original_text": "91 9820334455", "entity_value": { "value": "919820334455" }, "language": "en" } ] }
-
-
Example 2: Detecting phone number (hindi) from message
-
Django Shell:
message = u'मेरा मोबाइल नंबर है ९८९१९८९८७१' entity_name = 'phone_number' structured_value = None fallback_value = None bot_message = None source_langauge='hi' # here language will be ISO 639-1 code from ner_v2.detectors.pattern.phone_number.phone_number_detection import PhoneDetector detector = PhoneDetector(language=source_langauge, entity_name=entity_name) output = detector.detect(message=message, entity_name=entity_name, structured_value=structured_value, fallback_value=fallback_value, bot_message=bot_message,language=source_language) print(output)
-
CURL command:
curl -G -i "http://localhost:8081/v2/phone_number/?&entity_name=phone_number&source_language=hi" \ --data-urlencode "structured_value=" \ --data-urlencode "fallback_value=" \ --data-urlencode "bot_message=" \ --data-urlencode "message=मेरा मोबाइल नंबर है ९८९१९८९८७१"
**Output **:
{ "data": [ { "detection": "message", "original_text": "९८९१९८९८७१", "entity_value": { "value": "9891989871" }, "language": "hi" } ] }
-
-
Example 2: Detecting phone number from fallback value
-
Django Shell:
message = u'Please call me' entity_name = 'phone_number' structured_value = None fallback_value = '9049961794' bot_message = None source_langauge='en' from ner_v2.detectors.pattern.phone_number.phone_number_detection import PhoneDetector detector = PhoneDetector(language=source_langauge, entity_name=entity_name) output = detector.detect(message=message, entity_name=entity_name, structured_value=structured_value, fallback_value=fallback_value, bot_message=bot_message,language=source_language) print(output)
-
CURL command:
curl -G -i "http://localhost:8081/v2/phone_number/?&entity_name=phone_number&source_language=en" \ --data-urlencode "structured_value=" \ --data-urlencode "fallback_value=9049961794" \ --data-urlencode "bot_message=" \ --data-urlencode "message=Please call me"
**Output **:
{ "data": [ { "detection": "fallback_value", "original_text": "9049961794", "entity_value": { "value": "9049961794" }, "language": "en" } ] }
-
The Email Detector has the capability to detect emails within the given text.
API Example:
-
Example 1: Detecting emails from message
-
Django Shell:
message = u'my email id is [email protected]' entity_name = 'email' structured_value = None fallback_value = None bot_message = None from ner_v1.chatbot.entity_detection import get_email output = get_email(message=message,entity_name=entity_name, structured_value=structured_value, fallback_value=fallback_value, bot_message=bot_message) print(output)
-
CURL command:
curl -G -i "http://localhost:8081/v1/email/?&entity_name=email&source_language=en" \ --data-urlencode "structured_value=" \ --data-urlencode "fallback_value=" \ --data-urlencode "bot_message=" \ --data-urlencode "message=my email id is [email protected]"
**Output **
{ "data": [ { "detection": "message", "original_text": "[email protected]", "entity_value": { "value": "[email protected]" }, "language": "en" } ] }
-
-
Example 2: Detecting email from fallback value
-
Django Shell:
message = u'send this me to my email' entity_name = 'email' structured_value = None fallback_value = '[email protected]' bot_message = None from ner_v1.chatbot.entity_detection import get_email output = get_email(message=message,entity_name=entity_name, structured_value=structured_value, fallback_value=fallback_value, bot_message=bot_message) print(output)
-
CURL command:
curl -G -i "http://localhost:8081/v1/email/?&entity_name=email&source_language=en" \ --data-urlencode "structured_value=" \ --data-urlencode "[email protected]" \ --data-urlencode "bot_message=" \ --data-urlencode "message=send this me to my email"
**Output **
{ "data": [ { "detection": "fallback_value", "original_text": "[email protected]", "entity_value": { "value": "[email protected]" }, "language": "en" } ] }
-
-
The Text Detector has the capability to detect custom text entity within the given text. This detector is language agnostic
To create new text type entities, follow the steps from here.
API Examples:
-
Example 1: Detecting text entity from message
-
Django Shell:
message=u'i want to order chinese from mainland china and pizza from dominos' entity_name='restaurant' # here detection NER will search for values in dictionary 'restaurant' structured_value=None fallback_value=None bot_message=None source_language='en' from ner_v1.chatbot.entity_detection import get_text output = get_text(message=message, entity_name=entity_name, structured_value=structured_value, fallback_value=fallback_value, bot_message=bot_message,language=source_language) print(output)
CURL command:
curl -G -i "http://localhost:8081/v1/text/?&entity_name=restaurant&source_language=en" \ --data-urlencode "structured_value=" \ --data-urlencode "fallback_value=" \ --data-urlencode "bot_message=" \ --data-urlencode "message=i want to order chinese from mainland china and pizza from dominos"
**Output **:
{ "data": [ { "detection": "message", "original_text": "mainland china", "entity_value": { "crf_model_verified": false, "datastore_verified": true, "value": "Mainland China" }, "language": "en" }, { "detection": "message", "original_text": "dominos", "entity_value": { "crf_model_verified": false, "datastore_verified": true, "value": "Domino's Pizza" }, "language": "en" }, { "detection": "message", "original_text": "chinese", "entity_value": { "crf_model_verified": false, "datastore_verified": true, "value": "Yo! Chinese" }, "language": "en" }, { "detection": "message", "original_text": "pizza", "entity_value": { "crf_model_verified": false, "datastore_verified": true, "value": "U S Pizza" }, "language": "en" } ] }
-
-
Example 2: Detecting text entity from structured value
-
Django Shell:
message = u'मेरे लिए कैब बुक कर दीजिये' entity_name = 'city' structured_value = 'मुंबई' fallback_value = None bot_message = None source_langauge='hi' from ner_v1.chatbot.entity_detection import get_text output = get_text(message=message, entity_name=entity_name, structured_value=structured_value, fallback_value=fallback_value, bot_message=bot_message,langauge=source_language) print(output)
-
CURL command:
curl -G -i "http://localhost:8081/v1/text/?&entity_name=movie&source_language=hi" \ --data-urlencode "structured_value=मुंबई" \ --data-urlencode "fallback_value=" \ --data-urlencode "bot_message=" \ --data-urlencode "message=मेरे लिए कैब बुक कर दीजिये"
**Output **:
{ "data": [ { "detection": "structure_value_not_verified", "original_text": "मुंबई", "entity_value": { "value": "मुंबई" }, "language": "hi" } ] }
-
The PNR Detector has the capability to detect Train/ Flight PNR number within the given text.
API Examples:
-
Example 1: Detecting 10 digit Train PNR number from text
-
Django Shell
message = 'check my pnr status for 2141215305' entity_name = 'train_pnr' structured_value = None fallback_value = None bot_message = None from ner_v1.chatbot.entity_detection import get_pnr output = get_pnr(message=message, entity_name=entity_name, structured_value=structured_value, fallback_value=fallback_value, bot_message=bot_message) print(output)
-
CURL command:
curl -G -i "http://localhost:8081/v1/pnr/?&entity_name=pnr&source_language=en" \ --data-urlencode "structured_value=" \ --data-urlencode "fallback_value=" \ --data-urlencode "bot_message=" \ --data-urlencode "message=check my pnr status for 2141215305"
Output:
{ "data": [ { "detection": "message", "original_text": "2141215305", "entity_value": { "value": "2141215305" }, "language": "en" } ] }
-
Detect entities that match by the specified pattern. If you are not familiar with regex, please see http://www.rexegg.com/regex-quickstart.html
IMPORTANT NOTES
- The regex pattern provided must be escaped if you are not passing in a raw string (marked by 'r' in Python)
- Errors in compiling the provided pattern are not handled and will result in an exception
- chatbot_ner also uses re.UNICODE flag by default for detection. This can be overridden by using re_flags argument in the constructor
- If you are using groups, only 0th group will be returned. Sub grouping is not supported at the moment
API Examples:
-
Example 1: Detecting 4-6 digit number using regex
-
Django Shell:
message = 'please apply AMAZON30 coupon code to my cart' entity_name = 'regex_coupon_code' structured_value = None fallback_value = None bot_message = 'Enter the coupon code' regex = '[A-Z]+\d{2,6}' from ner_v1.chatbot.entity_detection import get_regex output = get_regex(message=message,entity_name=entity_name, structured_value=structured_value, fallback_value=fallback_value,bot_message=bot_message, pattern=regex) print(output)
-
CURL command:
curl -G -i "http://localhost:8081/v1/regex/?&entity_name=regex_coupon_code&source_language=en" \ --data-urlencode "structured_value=" \ --data-urlencode "fallback_value=" \ --data-urlencode "bot_message=" \ --data-urlencode "regex=[A-Z]+\d{2,6}" \ --data-urlencode "message=please apply AMAZON30 coupon code to my cart"
Output:
{ "data": [ { "detection": "message", "original_text": "AMAZON30", "entity_value": { "value": "AMAZON30" } } ] }
-