Retrieving Conversation Data

Overview

Hyperia can ingest any online meeting (Zoom, Google Meet, Microsoft Teams, ..) or media file (audio or video) for analysis, transcription, computer vision analysis, and NLP enrichment. A variety of ingestion mechanisms are provided including an AI notetaker bot that joins active meetings, direct file upload, and Internet web crawling (public http/https URLs). In addition to audio transcription, many other enrichments are provided including dialog act labeling, sentiment analysis, topic extraction, intent and action extraction, summarization, shared screen analysis (slide extraction, title extraction, screen OCR and text extraction, etc). APIs are provided to retrieve the various types of enrichment data provided by Hyperia.

Get Conversation Info

Retrieves document-level metadata for an ingested and analyzed conversation. Metadata includes conversation title and description, duration, and speakers.

Endpoint:

~~/v1/workspace/id/<workspace_id>/doc/id/<doc_id>/metadata/info~~

HTTP Method:

~~GET~~

URI Parameters:

Name	Description
workspace_id	ID of the workspace containing a document to retrieve a transcript from
doc_id	ID of the document to retrieve a transcript from

Returns:

If successful: ~~HTTP 200~~

Return Payload:

RESPONSE

{
  "status": "ok",
  "info": {
	"title": "DOCUMENT_TITLE",
	"description": "DOCUMENT_DESCRIPTION",
	"duration": CONVERSATION_DURATION,
	"speakers": [
	  {
		"id": "SPEAKER_GUID",
		"name": "SPEAKER_NAME"
	  }
	]
  }
}

Code Sample:

from hyperia import Hyperia
import sys

workspace_id = "SOME_WORKSPACE_ID"
doc_id = "SOME_DOCUMENT_ID"

hyperia = Hyperia()

response = hyperia.workspace_doc_metadata_info(workspace_id, doc_id)

if not response:
  	sys.exit(-1)

doc_data = response['info']

print(f"{doc_data['title']}: {doc_data['duration']}")

Get Transcript

Retrieves a transcription for an ingested and analyzed recording. Transcripts include speaker name and timestamp information.

Endpoint:

~~/v1/workspace/id/<workspace_id>/doc/id/<doc_id>/metadata/transcript~~

HTTP Method:

~~GET~~

URI Parameters:

Name	Description
workspace_id	ID of the workspace containing a document to retrieve a transcript from
doc_id	ID of the document to retrieve a transcript from

Returns:

If successful: ~~HTTP 200~~

Return Payload:

RESPONSE

{
    "status": "ok",
    "transcript": [
        {
		  "startTime": UTTERANCE_START_TIME_IN_SECONDS,
		  "endTime": UTTERANCE_END_TIME_IN_SECONDS,
		  "speaker": {
			"id": "SPEAKER_GUID",
			"name": "SPEAKER_NAME"
		  },
		  "transcript": "TRANSCRIPT_TEXT"
		}
	]
}

Code Sample:

from hyperia import Hyperia
import sys

workspace_id = "SOME_WORKSPACE_ID"
doc_id = "SOME_DOCUMENT_ID"

hyperia = Hyperia()

response = hyperia.workspace_doc_metadata_transcript(workspace_id, doc_id)

if not response:
  	sys.exit(-1)

transcript = response['transcript']

for utterance in transcript:
  print(f"{utterance['speaker']['name']}: {utterance['transcript']}")

Get Summary

Retrieves a machine-generated summary of an ingested and analyzed conversation. Summaries include timing and speaker information, as well as summarized key event data.

Endpoint:

~~/v1/workspace/id/<workspace_id>/doc/id/<doc_id>/metadata/summary~~

HTTP Method:

~~GET~~

URI Parameters:

Name	Description
workspace_id	ID of the workspace containing a document to retrieve summaries from
doc_id	ID of the document to retrieve summaries from

Returns:

If successful: ~~HTTP 200~~

Return Payload:

RESPONSE

{
    "status": "ok",
    "summary": [
        {
		  "startTime": SUMMARY_BLOCK_START_TIME_IN_SECONDS,
		  "endTime": SUMMARY_BLOCK_END_TIME_IN_SECONDS,
		  "speakers": [
			{
			  "id": "SPEAKER_GUID",
			  "name": "SPEAKER_NAME"
		    }
		  ],
		  "summaries": [
			{
			  "startTime": SUMMARY_ITEM_START_TIME_IN_SECONDS,
			  "endTime": SUMMARY_ITEM_END_TIME_IN_SECONDS,
			  "speakers": [
			    {
			       "id": "SPEAKER_GUID",
			       "name": "SPEAKER_NAME"
		        }
		      ],
			  "text": "SUMMARY_ITEM_TEXT"
			}
		  ]
		}
	]
}

Code Sample:

from hyperia import Hyperia
import sys

workspace_id = "SOME_WORKSPACE_ID"
doc_id = "SOME_DOCUMENT_ID"

hyperia = Hyperia()

response = hyperia.workspace_doc_metadata_summary(workspace_id, doc_id)

if not response:
  	sys.exit(-1)

summary = response['summary']
for block in summary['summaries']:
  print(block)

Get Monologues

Retrieves a machine-generated monologues for an ingested and analyzed conversation. Monologues include timing, title, and speaker information, as well as transcript data.

Endpoint:

~~/v1/workspace/id/<workspace_id>/doc/id/<doc_id>/metadata/monologues~~

HTTP Method:

~~GET~~

URI Parameters:

Name	Description
workspace_id	ID of the workspace containing a document to retrieve monologues from
doc_id	ID of the document to retrieve monologues from

Returns:

If successful: ~~HTTP 200~~

Return Payload:

RESPONSE

{
    "status": "ok",
    "monologues": [
        {
		  "startTime": MONOLOGUE_START_TIME_IN_SECONDS,
		  "endTime": MONOLOGUE_END_TIME_IN_SECONDS,
		  "speakers": [
			{
			  "id": "SPEAKER_GUID",
			  "name": "SPEAKER_NAME"
		    }
		  ],
		  "title": "MONOLOGUE_TITLE",
		  "transcript": [
			{
			  "startTime": UTTERANCE_START_TIME_IN_SECONDS,
			  "endTime": UTTERANCE_END_TIME_IN_SECONDS,
			  "speaker": {
				"id": "SPEAKER_GUID",
				"name": "SPEAKER_NAME"
			  },
			  "transcript": "TRANSCRIPT_TEXT"
			}
		  ]
		}
	]
}

Code Sample:

from hyperia import Hyperia
import sys

workspace_id = "SOME_WORKSPACE_ID"
doc_id = "SOME_DOCUMENT_ID"

hyperia = Hyperia()

response = hyperia.workspace_doc_metadata_monologues(workspace_id, doc_id)

if not response:
  	sys.exit(-1)

monologues = response['monologues']
for monologue in monologues:
  print(monologue)

Get Screenshots

Retrieves automatically extracted screenshot data of any content display during an ingested and analyzed conversation. Screen data includes extracted OCR information, bounding boxes, and timing information.

Endpoint:

~~/v1/workspace/id/<workspace_id>/doc/id/<doc_id>/metadata/screens~~

HTTP Method:

~~GET~~

URI Parameters:

Name	Description
workspace_id	ID of the workspace containing the document to retrieve screens from
doc_id	ID of the document to retrieve screens from

Returns:

If successful: ~~HTTP 200~~

Return Payload:

RESPONSE

{
    "status": "ok",
    "screens": [
        {
		  "startTime": SCREEN_VISIBILITY_START_TIME_IN_SECONDS,
		  "endTime": SCREEN_VISIBILITY_END_TIME_IN_SECONDS,
		  "title": "OCR_SCREEN_TITLE_DATA",
		  "ocr": {
			"lines": [
			  {
				"text": "OCR_LINE_TEXT_DATA"
				"minX": BOUNDING_BOX_MIN_X,
				"minY": BOUNDING_BOX_MIN_Y,
				"maxX": BOUNDING_BOX_MAX_X,
				"maxY": BOUNDING_BOX_MAX_Y
			  }
		  	]
		  },
		  "mentions": [
			{
			  "startTime": UTTERANCE_START_TIME_IN_SECONDS,
			  "endTime": UTTERANCE_END_TIME_IN_SECONDS,
			  "speaker": {
				"id": "SPEAKER_GUID",
				"name": "SPEAKER_NAME"
			  },
			  "transcript": "TRANSCRIPT_TEXT"
			}
		  ]
		}
	]
}

Code Sample:

from hyperia import Hyperia
import sys

workspace_id = "SOME_WORKSPACE_ID"
doc_id = "SOME_DOCUMENT_ID"

hyperia = Hyperia()

response = hyperia.workspace_doc_metadata_screens(workspace_id, doc_id)

if not response:
  	sys.exit(-1)

screens = response['screens']
for screen in screens:
  print(screen)

Get Extracted Topics

Retrieves automatically extracted natural language topics for any an ingested and analyzed conversation. Topic data includes topic name and type, speaker and timing information.

Endpoint:

~~/v1/workspace/id/<workspace_id>/doc/id/<doc_id>/metadata/topics~~

HTTP Method:

~~GET~~

URI Parameters:

Name	Description
workspace_id	ID of the workspace containing the document to retrieve topics from
doc_id	ID of the document to retrieve topics from

Returns:

If successful: ~~HTTP 200~~

Return Payload:

RESPONSE

{
    "status": "ok",
    "topics": [
        {
		  "text": "TOPIC_NAME",
		  "type": "TOPIC_TYPE",
		  "count": OCCURRENCE_COUNT,
		  "mentions": [
			{
			  "startTime": UTTERANCE_START_TIME_IN_SECONDS,
			  "endTime": UTTERANCE_END_TIME_IN_SECONDS,
			  "speaker": {
				"id": "SPEAKER_GUID",
				"name": "SPEAKER_NAME"
			  },
			  "transcript": "TRANSCRIPT_TEXT"
			}
		  ]
		}
	]
}

Code Sample:

from hyperia import Hyperia
import sys

workspace_id = "SOME_WORKSPACE_ID"
doc_id = "SOME_DOCUMENT_ID"

hyperia = Hyperia()

response = hyperia.workspace_doc_metadata_topics(workspace_id, doc_id)

if not response:
  	sys.exit(-1)

topics = response['topics']
for topic in topics:
  print(topic)

Get Automation Tags

Retrieves tags created within an ingested and analyzed conversation. Tags could be created by a Hyperia user during the process of creating or sharing a comment or clip, or automatically by a Hyperia Automation.

Endpoint:

~~/v1/workspace/id/<workspace_id>/doc/id/<doc_id>/metadata/tags~~

HTTP Method:

~~GET~~

URI Parameters:

Name	Description
workspace_id	ID of the workspace containing the document to retrieve tags from
doc_id	ID of the document to retrieve tags from

Returns:

If successful: ~~HTTP 200~~

Return Payload:

RESPONSE

{
    "status": "ok",
    "tags": [
        {
		  "name": "TAG_NAME",
		  "id": "TAG_GUID",
		  "count": OCCURRENCE_COUNT,
		  "mentions": [
			{
			  "startTime": UTTERANCE_START_TIME_IN_SECONDS,
			  "endTime": UTTERANCE_END_TIME_IN_SECONDS,
			  "speaker": {
				"id": "SPEAKER_GUID",
				"name": "SPEAKER_NAME"
			  },
			  "transcript": "TRANSCRIPT_TEXT"
			}
		  ]
		}
	]
}

Code Sample:

from hyperia import Hyperia
import sys

workspace_id = "SOME_WORKSPACE_ID"
doc_id = "SOME_DOCUMENT_ID"

hyperia = Hyperia()

response = hyperia.workspace_doc_metadata_topics(workspace_id, doc_id)

if not response:
  	sys.exit(-1)

tags = response['tags']
for tag in tags:
  print(tag)

Cookies 🍪

Retrieving Conversation Data

Overview

Get Conversation Info

Get Transcript

Get Summary

Get Monologues

Get Screenshots

Get Extracted Topics

Get Automation Tags