emergingtrajectories¶

Submodules¶

Classes¶

`Statement`
`Forecast`
`Client`	Launch the Emerging Trajectories Client.
`EmergingTrajectoriesClient`

Functions¶

hello(→ None)

Just a hello() message/function to confirm you've installed everything!

Package Contents¶

emergingtrajectories.hello() → None¶: Just a hello() message/function to confirm you’ve installed everything!

class emergingtrajectories.Statement(title, fill_in_the_blank)¶

Bases: object

id = -1¶

title¶

fill_in_the_blank¶

description = ''¶

deadline = None¶

created_at = None¶

updated_at = None¶

created_by = None¶

class emergingtrajectories.Forecast(title, value, justification)¶

Bases: object

id = -1¶

title¶

value¶

justification¶

statement = None¶

created_at = None¶

updated_at = None¶

created_by = None¶

prediction_agent = None¶

additional_data¶

prior_forecast = None¶

next_forecasts = []¶

is_human = False¶

class emergingtrajectories.Client(api_key: str)¶

Bases: object

Launch the Emerging Trajectories Client.

Parameters:: api_key – the API key for the Emerging Trajectories platform.

base_url = 'https://app.emergingtrajectories.com/a/api/v0.2/'¶

api_key¶

create_statement(title: str, description: str, deadline: datetime.datetime, fill_in_the_blank: str) → dict¶

Create a new statement that users will be forecasting against/for.

Parameters:

title – the title of the statement
description – a more detailed description of the statement
deadline – the deadline (date and time), typically when we’ll learn the “right answer” for the forecasting process
fill_in_the_blank – the “fill in the blank” part of the statement, which is what users will be forecasting against

Returns:

the newly created statement returned from the platform

Return type:

dict

get_statement(statement_id: int) → dict¶

Returns a given statement from the platform. Includes title, description, deadline, and fill-in-the-blank.

Parameters:: statement_id – the ID of the statement to retrieve
Returns:: the statement from the platform
Return type:: dict

get_most_recent_forecast(statement_id: int, prediction_agent: str = None) → int¶

Returns the most recent forecast for a given statement. This is useful for creating a new forecast that is an extension of a prior forecast.

Parameters:

statement_id – the ID of the statement to retrieve the most recent forecast for
prediction_agent – the string for a prediction agent, if you want to further filter the most recent forecast

Returns:

the ID of the most recent forecast for the given statement

Return type:

int

get_forecast(forecast_id: int) → dict¶

Returns a specific forecast’s details from the platform. This typically includes the forecast title, the value associated with the fill-in-the-blank component of a statement, and justificaiton for the forecast.

Parameters:: forecast_id – the ID of the forecast to retrieve
Returns:: the forecast details in the form of a dictionary object
Return type:: dict

create_forecast(statement_id: int, title: str, justification: str, value: float, prediction_agent: str, additional_data: dict = {}, prior_forecast: int = None, is_human: bool = False) → None¶

Creates a forecast tied to a specific statement.

Parameters:

statement_id – the ID of the statement to tie the forecast to
title – the title of the forecast
justification – the justification for the forecast
value – the value associated with the fill-in-the-blank component of the statement
prediction_agent – the agent making the prediction
additional_data – any additional data to include with the forecast. This is not used anywhere, but can be helpful in audting or researching forecast effectiveness
prior_forecast – if this forecast is an extension of an earlier forecast, the ID of the prior forecast
is_human – whether the prediction is human-generated

Returns:

the newly created forecast from the platform

Return type:

dict

class emergingtrajectories.EmergingTrajectoriesClient(api_key: str)¶

Bases: object

Parameters:: api_key – the API key for the Emerging Trajectories platform.

base_url = 'https://app.emergingtrajectories.com/a/api/v0.2/'¶

api_key¶

get_events(project_codes: str | List[str] = None, limit: int = 100) → List[dict]¶

Get events from the Emerging Trajectories platform.

Parameters:: project_codes – a single project code or a list of project codes to filter events by
Returns:: a list of events, each represented as a dictionary
Return type:: List[dict]

search_public_events(query, project_codes: str | List[str], limit: int = 10) → List[dict]¶

Search for public events on the Emerging Trajectories platform.

Parameters:

query – the search query string
project_codes – a single project code or a list of project codes to filter events by
limit – the maximum number of events to return

Returns:

a list of events, each represented as a dictionary

Return type:

List[dict]

upload_pdf(pdf_path: str, factbase_shortcode: str) → bool¶

Upload a PDF to the Emerging Trajectories platform.

Parameters:

pdf_path – the path to the PDF file to upload
factbase_shortcode – the short code for the factbase to attach the PDF to

Returns:

returns True if file has been uploaded successfully

Return type:

bool

create_factbase(title: str, description: str) → str¶

Create a new factbase on the Emerging Trajectories platform.

Parameters:

title – the title of the factbase
description – a description of the factbase

Returns:

The short_code for the factbase.

Return type:

str

get_facts(short_code, query, days_filter=1, pubdate_days_filter=-1, use_text_match=False) → str¶

Query the fact base and simply get the raw facts normally generated for documents or other internal processes.

Parameters:

short_code – the short code for the factbase to query
query – the query to run
days_filter – limit facts to the recent number of days
pubdate_days_filter – limit facts to the recent number of days based on publication date
use_text_match – use text matching for the query

Returns:

A list of facts as a string.

Return type:

str

query_factbase(short_code, query, days_filter=1, pubdate_days_filter=-1, llm_model=None, llm_temperature=None) → str¶

Query the fact base as if you’d query a document, but with no document required. The responses here are ephemeral and are not stored anywhere.

Parameters:

short_code – the short code for the factbase to query
query – the query to run
days_filter – limit facts to the recent number of days
pubdate_days_filter – limit facts to the recent number of days based on publication date
llm_model – the LLM model to use
llm_temperature – the LLM temperature to use

Returns:

The response written based on facts in the fact base.

Return type:

str

build_research_plan(query: str) → dict¶

Given a research question/query, we create a set of tasks that we can then automate document creation around.

Parameters:: query – the research question/query to build tasks around
Returns:: a dictionary with two keys – “text” which is the text plan, and “plan” which is a JSON plan
Return type:: dict

run_research_sub_task(command: str, args: str | List[str], document_id: int, short_code: str, wait_until_completion: bool = True) → int¶

Run the research sub-task outlined in a research plan.

Parameters:

command – the command to run
args – the arguments to pass to the command
document_id – the ID of the document to attach the research task to
short_code – the short code for the factbase to use
wait_until_completion – whether to wait until the task is completed before returning

Returns:

the ID of the job or -1 if not applicable

Return type:

int

research_task_header_and_block(doc_id: int, query: str) → int¶

Runs a research analyst task where we generate a header block and then run the query for the document.

Parameters:

doc_id – the ID of the document to attach the research task to
query – the query to run

Returns:

the ID of the job

Return type:

int

run_data_collector_job(factbase_shortcode: str, settings: dict) → int¶

Creates a new data collector job and runs it. Contact us for information on how to pass settings.

Parameters:

factbase_shortcode – the short code for the factbase
settings – a dictionary of settings for the data collector job

Returns:

The ID of the job

Return type:

int

run_data_collector_serp(factbase_shortcode: str, query: str | List[str], n: int = 5, settings: dict = None) → int¶

Creates a new data collector job and runs it. Contact us for information on how to pass settings.

Parameters:

factbase_shortcode – the short code for the factbase
query – a string to search or an array of strings
settings – a dictionary of settings for the data collector job

Returns:

The ID of the job

Return type:

int

run_data_collector_news(factbase_shortcode: str, query: str | List[str], n: int = 5, settings: dict = None) → int¶

Creates a new data collector job and runs it. Contact us for information on how to pass settings.

Parameters:

factbase_shortcode – the short code for the factbase
query – a string to search or an array of strings
settings – a dictionary of settings for the data collector job

Returns:

The ID of the job

Return type:

int

get_data_collector_job_status(job_id: int) → str¶

Get the status of a data collector job.

Parameters:: job_id – the ID of the job
Returns:: The status of the job
Return type:: str

create_document(factbase_shortcode: str, title: str = None, days_filter=99999, temperature=None, llm=None, facts_min_date=None, facts_max_date=None) → int¶

Create a new document on the Emerging Trajectories platform.

Parameters:

factbase_shortcode – the short code for the factbase to attach the document to
title (optional) – the title of the document
days_filter (optional) – limit facts to the recent number of days

Returns:

The ID of the document.

Return type:

int

convert_facts_in_text(factbase_shortcode: str, text: str)¶

Convert text with facts to HTML code with associated links.

Parameters:

factbase_shortcode – the short code for the factbase to attach the document to
text – the text to convert

Returns:

the converted text as HTML array: a list facts found in the text, with the key being the source ID in the HTML

Return type:

str

update_document(doc_id: int, title: str = None, is_public: bool = None, short_code: str = None) → bool¶

Update document settings.

Parameters:

doc_id – the ID of the document
title – the new title of the document
is_public – whether the document is public (updated setting)
short_code – the short code for the document (updated setting)

Returns:

True if successful, False otherwise

Return type:

bool

get_document(doc_id: int) → dict¶

Get document data.

Parameters:: doc_id – the ID of the document
Returns:: The ID of the document.
Return type:: int

add_viewer(doc_id: int, viewer_email: str) → bool¶

Add a viewer to a private document.

Parameters:

doc_id – the ID of the document
viewer_email – the email of the viewer to add

Returns:

True if successful, False otherwise

Return type:

bool

remove_viewer(doc_id: int, viewer_email: str) → bool¶

Remove a viewer from a private document.

Parameters:

doc_id – the ID of the document
viewer_email – the email of the viewer to remove

Returns:

True if successful or if the email is NOT a viewer already, False otherwise

Return type:

bool

get_block(doc_id: int, block_named_id: str) → str¶

Get the content of a named block from a document.

Parameters:

doc_id – the ID of the document
block_named_id – the named ID of the block to retrieve

Returns:

the content of the block

Return type:

str

append_header_block(doc_id, text: str, is_hidden: bool = False) → bool¶

Add a header block to a document.

Parameters:

doc_id – the ID of the document to append the header block to
text – the text (header) to append
is_hidden – whether the block is hidden in public documents

Returns:

True if successful, False otherwise

Return type:

bool

append_text_block(doc_id: int, text: str) → bool¶

Add a text block to a document.

Parameters:

doc_id – the ID of the document to append the text block to
text – the text to append

Returns:

True if successful, False otherwise

Return type:

bool

append_hidden_block(doc_id: int, prompt: str) → bool¶

Add a text block to a document.

Parameters:

doc_id – the ID of the document to append the text block to
prompt – the text (typically a prompt) to append

Returns:

True if successful, False otherwise

Return type:

bool

append_askai_fts_block(doc_id: int, query: str, fts_term: str, named_id: str = None, is_hidden: bool = False) → dict¶

Add an AI (FTS) block to a document.

Parameters:

doc_id – the ID of the document to append the text block to
query – the query to use for the AI block
fts_term – the full text search (FTS) terms/query to use for the AI block
named_id – the named ID of the AI block
is_hidden – whether the block is hidden in public documents

Returns: