pipebridge 0.3.0

Semantic integration layer for building scalable and reliable Pipefy automations.

PipeBridge

PipeBridge is a Python SDK for Pipefy that gives you a semantic, reliable layer for building production-grade automations.

Instead of wiring raw GraphQL queries, manual validation, and brittle payload handling into every integration, PipeBridge gives you predictable workflows, typed models, and extension points that fit real automation scenarios.

PipeBridge is not a thin GraphQL wrapper. It is an integration framework designed for maintainable Pipefy automation.

New in v0.3.0:

• first-class connector discovery and semantic connector operations
• connector-safe start form creation and card updates
• richer table-backed connector options with record_fields and record_fields_map
• phase-schema helpers for explicit field existence checks
• file flows aligned with schema-based attachment validation instead of card.fields

Quick links:

• Documentation: https://rmcavalcante7.github.io/pipebridge/
• Use cases: https://github.com/rmcavalcante7/pipebridge/tree/main/useCases
• PyPI: https://pypi.org/project/pipebridge/

Summary

• Why PipeBridge?
• Installation
• Quick Start
• Public Surface
• Core Capabilities
• Start Form Semantics
• Full Structure Traversal
• Extensibility
• Models and Semantic Navigation
• Transport Configuration
• Schema Cache
• Ready-to-Use Examples
• HTML Documentation
• Tests
• Current Status
• Vision
• Author
• License

Why PipeBridge?

Direct Pipefy integrations usually force you to deal with:

• verbose GraphQL operations
• inconsistent payload structures
• repeated validation logic
• unsafe phase transitions
• ad hoc file handling and retries

PipeBridge addresses that with:

• a simple public facade
• typed models with semantic navigation
• start form-aware schema discovery
• safe card creation
• safe field updates
• safe phase moves
• file upload and download flows
• transport-level TLS and retry controls
• schema caching
• extensibility through rules, handlers, policies, and steps

Installation

pip install pipebridge

For development:

pip install -e .[dev]

Quick Start

from pipebridge import PipeBridge

api = PipeBridge(
  token="YOUR_TOKEN",
  base_url="https://app.pipefy.com/queries",
)

card = api.cards.get("123456789")
print(card.title)
print(card.current_phase.name if card.current_phase else None)

Public Surface

The main SDK entry point is the facade:

api = PipeBridge(token="YOUR_TOKEN", base_url="https://app.pipefy.com/queries")

Public domains:

• api.cards
• api.phases
• api.pipes
• api.connectors
• api.files

Sub-levels when applicable:

• api.cards.raw
• api.cards.structured
• api.phases.raw
• api.phases.structured
• api.pipes.raw
• api.pipes.structured

Objects also exposed at the package top level:

• PipefyHttpClient
• TransportConfig
• CardService
• FileService
• PipeService
• PhaseService
• ConnectorService
• FileUploadRequest
• FileDownloadRequest
• UploadConfig
• CardUpdateConfig
• CardMoveConfig

Core Capabilities

1. Card, pipe, and phase retrieval

card = api.cards.get("123")
phase = api.phases.get("456")
pipe = api.pipes.get("789")

Important note about card field payloads:

• card.fields mirrors the fields collection returned by Pipefy for that card query
• it should be treated as the set of materialized field values exposed by the API, not as the complete schema of the phase or pipe
• a field missing from card.fields does not imply that the field does not exist in the phase or pipe schema
• for schema existence checks, prefer phase or pipe helpers

2. Pipe schema catalog and start form coverage

pipe = api.pipes.getFieldCatalog("789")

for field in pipe.iterStartFormFields():
    print(field.id, field.internal_id, field.type, field.required)

for phase in pipe.iterPhases():
    print(phase.name)
    for field in phase.iterFields():
        print(field.id, field.internal_id, field.type, field.options)

This catalog is important for:

• field discovery
• start form discovery
• update validation
• type support
• schema caching
• internal field mapping via internal_id

3. Safe card creation from start form

result = api.cards.createSafely(
  pipe_id="789",
  title="New request",
  fields={
    "oc": "12345",
    "request_type": "Purchase",
  },
)

This path is intended for users who prefer a more conservative entry flow.

It validates:

• whether the field belongs to the pipe start form
• whether required start form fields were filled
• whether option values are valid when the field exposes options

Important note:

• start-form connector fields must receive connected record ids, not display labels
• api.connectors.resolveOption(...) can be used before creation when the caller only knows the display title
• a full tenant-specific example of create, move, and update is available in useCases/start_form_create_move_fill.py

4. Connector discovery and semantic operations

fields = api.connectors.listFields("789")

for connector in fields:
    print(
        connector.field_id,
        connector.origin_type,
        connector.phase_name,
        connector.connected_repo.repo_type if connector.connected_repo else None,
        connector.connected_repo.name if connector.connected_repo else None,
    )

option = api.connectors.resolveOption(
    pipe_id="789",
    field_id="nome_projetos",
    title="IA Time",
)

api.connectors.setCardValue(
    card_id="123",
    field_id="nome_projetos",
    item_ids=[option.id],
)

Connector notes:

• connector options are dynamic and repo-backed
• connectors may live in the start form or in a regular phase
• table-backed connector options may expose extra record metadata through record_fields and record_fields_map
• that metadata can be used to disambiguate options that share similar titles
• empty connectors may be absent from card.fields
• for filled connectors, api.connectors.getCardValue(...) exposes ids and connected items

5. Safe card field updates

from pipebridge import CardUpdateConfig

result = api.cards.updateFields(
  card_id="123",
  fields={
    "title": "New value",
    "priority": "High",
  },
  expected_phase_id="456",
  config=CardUpdateConfig(
    validate_field_existence=True,
    validate_field_options=True,
    validate_field_type=True,
    validate_field_format=True,
  ),
)

The current update flow supports important field families, including:

• short and long text
• number
• currency
• email
• date
• datetime
• due_date
• time
• select
• radio
• label_select
• checklist
• assignee_select
• connector
• attachment

For explicit phase-schema inspection, the public phase helpers are:

field = api.phases.getField("456", "priority")
exists = api.phases.hasField("456", "priority")
required_field = api.phases.requireField("456", "priority")

These helpers answer whether a field exists in the phase schema, not whether a value was materialized in card.fields.

6. Safe phase moves

from pipebridge import CardMoveConfig

result = api.cards.moveSafely(
  card_id="123",
  destination_phase_id="999",
  expected_current_phase_id="456",
  config=CardMoveConfig(validate_required_fields=True),
)

This flow validates:

• whether the current phase matches the expected phase, when provided
• whether the transition is allowed by the current phase configuration
• whether required fields in the destination phase are filled

7. File upload and download

from pipebridge import FileUploadRequest, FileDownloadRequest, UploadConfig

upload_request = FileUploadRequest(
  file_name="sample.txt",
  file_bytes=b"content",
  card_id="123",
  field_id="attachments",
  organization_id="999",
  expected_phase_id="456",
)

upload_result = api.files.uploadFile(upload_request)

download_request = FileDownloadRequest(
  card_id="123",
  field_id="attachments",
  output_dir="./downloads",
)

files = api.files.downloadAllAttachments(download_request)

Important file-flow notes:

• attachment field existence is validated against schema, not only against the current card.fields payload
• attachment merge/download now uses the card attachment surface instead of assuming the attachment field must already be materialized in card.fields
• by default, upload validates attachment fields against the current phase schema
• when needed, UploadConfig(validate_field_in_current_phase=False) allows upload to attachment fields that exist elsewhere in the pipe schema, such as start form or another phase

Example:

upload_result = api.files.uploadFile(
  request=upload_request,
  config=UploadConfig(validate_field_in_current_phase=False),
)

8. Transport configuration

from pipebridge import PipeBridge, TransportConfig

api = PipeBridge(
  token="YOUR_TOKEN",
  base_url="https://app.pipefy.com/queries",
  transport_config=TransportConfig(
    timeout=45,
    verify_ssl=True,
    max_retries=2,
    retry_delay_seconds=1.0,
    retry_backoff_multiplier=2.0,
  ),
)

This transport layer supports:

• global request timeout override
• TLS verification control
• custom CA bundle path for corporate environments
• conservative retry for transient timeout and connection errors

Start Form Semantics

PipeBridge now models the start form as part of the pipe schema, but not as a regular phase.

That distinction matters because:

• Pipefy exposes start_form_fields at the pipe level
• card creation enters the pipe through the start form
• later navigation and movement still happen through regular phases

This keeps the SDK aligned with the platform instead of introducing a fake phase abstraction.

Full Structure Traversal

For a complete real example, see:

• useCases/pipe_cascade_inspection.py

Simplified loop:

pipe = api.pipes.get("PIPE_ID")

print(f"Pipe: {pipe.name} ({pipe.id})")

for phase_summary in pipe.iterPhases():
    phase = api.phases.get(phase_summary.id)
    print(f"Phase: {phase.name} ({phase.id})")

    for field in phase.iterFields():
        print(
            f"id={field.id} | "
            f"type={field.type} | "
            f"required={field.required} | "
            f"options={field.options}"
        )

    cards = api.phases.listCards(phase.id)
    for card in cards:
        print(f"Card: {card.title} ({card.id})")

        for card_field in card.iterFields():
            print(
                f"field_id={card_field.id} | "
                f"label={card_field.label} | "
                f"type={card_field.type} | "
                f"value={card_field.value}"
            )

Extensibility

One of the project's core goals is to allow extension without forking the SDK.

1. Custom rules

You can inject extra rules into public flows.

Example with updates:

from pipebridge.exceptions import ValidationError
from pipebridge.workflow.rules.baseRule import BaseRule


class UppercaseOnlyRule(BaseRule):
  def __init__(self, field_id: str) -> None:
    self.field_id = field_id

  def execute(self, context) -> None:
    value = context.request.fields.get(self.field_id)
    if not isinstance(value, str) or value != value.upper():
      raise ValidationError(
        message=f"Field '{self.field_id}' must be uppercase",
        class_name=self.__class__.__name__,
        method_name="execute",
      )


api.cards.updateField(
  card_id="123",
  field_id="code",
  value="VALUE",
  extra_rules=[UppercaseOnlyRule("code")],
)

2. Ready-to-use regex for field validation

from pipebridge.service.card.flows.update.rules.regexFieldPatternRule import (
  RegexFieldPatternRule,
)

api.cards.updateField(
  card_id="123",
  field_id="code",
  value="ABC-123",
  extra_rules=[
    RegexFieldPatternRule({"code": r"^[A-Z]{3}-\d{3}$"})
  ],
)

3. Custom update handlers

You can override or add type support at runtime:

from pipebridge.service.card.flows.update.dispatcher.baseCardFieldUpdateHandler import (
  BaseCardFieldUpdateHandler,
)
from pipebridge.service.card.flows.update.dispatcher.resolvedFieldUpdate import (
  ResolvedFieldUpdate,
)


class UppercaseTextHandler(BaseCardFieldUpdateHandler):
  def resolve(self, field_id, field_type, input_value, current_field=None, phase_field=None):
    return ResolvedFieldUpdate(
      field_id=field_id,
      field_type=field_type,
      input_value=input_value,
      current_field=current_field,
      phase_field=phase_field,
      new_value=str(input_value).strip().upper(),
    )


api.cards.updateField(
  card_id="123",
  field_id="title",
  value="my text",
  extra_handlers={"short_text": UppercaseTextHandler()},
)

4. Retry and circuit breaker policies

from pipebridge import UploadConfig
from pipebridge.workflow.config.retryConfig import RetryConfig
from pipebridge.workflow.config.circuitBreakerConfig import CircuitBreakerConfig

config = UploadConfig(
  retry=RetryConfig(max_retries=5, base_delay=1.0),
  circuit=CircuitBreakerConfig(failure_threshold=5, recovery_timeout=5.0),
)

api.files.uploadFile(request=upload_request, config=config)

5. Custom upload steps

In V1, steps extensibility is publicly exposed only for uploads:

• extra_steps_before
• extra_steps_after

from pipebridge.workflow.steps.baseStep import BaseStep


class RegisterMetadataStep(BaseStep):
  def execute(self, context) -> None:
    context.metadata["source"] = "custom-step"


api.files.uploadFile(
  request=upload_request,
  extra_steps_before=[RegisterMetadataStep()],
)

Note:

• card updates and safe moves do not yet expose custom steps in the V1 public API

Models and Semantic Navigation

SDK models were designed for semantic navigation. The goal is to avoid direct structural map access whenever possible.

Examples:

card = api.cards.get("123")

title_value = card.getFieldValue("title")
if title_value is not None:
    print(title_value)

phase = api.phases.get("456")
print(phase.getFieldType("priority"))
print(phase.getFieldOptions("priority"))
print(phase.isFieldRequired("priority"))

if api.phases.hasField("456", "priority"):
    schema_field = api.phases.requireField("456", "priority")
    print(schema_field.label, schema_field.type)

pipe = api.pipes.getFieldCatalog("789")
for field in pipe.getFieldsByType("select"):
    print(field.id, field.label)

for start_form_field in pipe.iterStartFormFields():
    print(start_form_field.id, start_form_field.internal_id)

for connector in api.connectors.listFields("789"):
    print(
        connector.field_id,
        connector.origin_type,
        connector.connected_repo.name if connector.connected_repo else None,
    )

Important semantic distinction:

• card.hasField(...) and card.getField(...) operate on the card.fields payload returned by Pipefy for that card query
• they answer whether the API materialized a value for that field in the card payload
• they do not answer whether a field exists in a phase schema or anywhere in the pipe
• for schema-oriented checks, use:
  • api.phases.hasField(phase_id, field_id)
  • api.phases.getField(phase_id, field_id)
  • api.phases.requireField(phase_id, field_id)
  • api.pipes.getFieldCatalog(pipe_id)

For connector options backed by tables, the returned option objects may also include:

• record_fields
• record_fields_map

This lets callers inspect extra identification attributes such as project manager, squad leader email, or responsible owner before connecting an item.

Transport Configuration

TransportConfig is the public transport-layer configuration object exposed at the top level of the package.

Use it when you need:

• timeout control across SDK operations
• custom certificate bundles in corporate networks
• temporary TLS relaxation in controlled environments
• bounded retries for transient transport failures

from pipebridge import PipefyHttpClient, TransportConfig

client = PipefyHttpClient(
  auth_key="YOUR_TOKEN",
  base_url="https://app.pipefy.com/queries",
  transport_config=TransportConfig(
    timeout=30,
    ca_bundle_path="/path/to/company-ca.pem",
    max_retries=1,
  ),
)

Schema Cache

The SDK provides in-memory cache for pipe schema:

• keyed by pipe_id
• with TTL
• with per-key locking
• lazy refresh on demand
• no background thread in V1

On the card facade:

stats = api.cards.getSchemaCacheStats()
entry = api.cards.getSchemaCacheEntryInfo("789")
api.cards.invalidateSchemaCache("789")

And for direct schema inspection:

pipe_schema = api.pipes.getFieldCatalog("789")

for phase in pipe_schema.iterPhases():
    print(f"Phase: {phase.name} ({phase.id})")
    for field in phase.iterFields():
        print(
            f"id={field.id} | "
            f"label={field.label} | "
            f"type={field.type} | "
            f"required={field.required}"
        )

Ready-to-Use Examples

The useCases folder is the recommended starting point for end users.

It contains executable examples for:

• pipe field catalog inspection
• cascading inspection across pipes, phases, and cards
• connector discovery, option resolution, and semantic updates
• start form creation followed by safe move and phase filling
• card field updates
• updates with extra rules
• custom handler
• safe moves
• upload and download
• uploads with rules and policies
• uploads with custom steps

See useCases/README.md.

HTML Documentation

The project also includes a Sphinx documentation structure in docs/.

This is the intended path for the SDK's navigable HTML documentation, including:

• overview
• quick start
• extensibility
• API reference
• development guides

To generate locally:

pip install -e .[docs]
sphinx-build -b html docs docs/_build/html

Main documentation entry point in the repository:

• docs/index.rst

Expected URL for published documentation via GitHub Pages:

• https://rmcavalcante7.github.io/pipebridge/

Tests

The project is organized as follows:

• tests/unit
• tests/functional
• tests/integration
• useCases/

Role of each:

• unit

  • isolated logic
  • no network
  • no credentials

• functional

  • public API
  • no real Pipefy
  • with fakes/doubles

• integration

  • real Pipefy operations
  • depend on:
    • PIPEFY_API_TOKEN
    • optional PIPEFY_BASE_URL

Commands:

python -m pytest tests/unit tests/functional -v
python -m pytest tests/integration -v
python -m pytest tests -v

For real integration:

$env:PIPEFY_API_TOKEN="YOUR_TOKEN"
$env:PIPEFY_BASE_URL="https://app.pipefy.com/queries"
python -m pytest tests/integration -v

For the destructive live create/move/update battery:

$env:PIPEFY_API_TOKEN="YOUR_TOKEN"
$env:PIPEFY_BASE_URL="https://app.pipefy.com/queries"
$env:PIPEBRIDGE_ENABLE_DESTRUCTIVE_CREATE_TESTS="1"
$env:PIPEBRIDGE_TEST_PIPE_ID="307064875"
$env:PIPEBRIDGE_REFERENCE_CARD_ID="1330664077"  # optional, read-only reference
$env:PIPEBRIDGE_DELETE_CREATED_TEST_CARD="1"    # default behavior
python -m pytest tests/integration/test_card_service.py tests/integration/test_card_move_flow.py tests/integration/test_card_update_flow.py -v

Notes for the destructive live battery:

• it creates one new card per test session
• all live mutations in that battery run only against the created card
• the optional reference card is read-only and is used only to copy values when helpful
• when no reference card is provided, the helpers generate valid values by field type
• teardown can delete only the created card when PIPEBRIDGE_DELETE_CREATED_TEST_CARD is enabled

Current Status

Current release highlights:

• coherent public facade
• start form-aware pipe schema catalog
• safe card creation via createSafely(...)
• transport configuration via TransportConfig
• connector schema discovery via api.connectors
• semantic connector read and update helpers
• TLS and retry controls at the HTTP boundary
• card update flow
• safe move flow
• upload/download flow
• semantic exceptions
• schema cache
• structured pytest suite
• end-user usage examples
• destructive live test battery that creates, updates, moves, validates, and optionally deletes only the card created for that run

Foundational capabilities already in place:

• coherent public facade
• card update flow
• safe move flow
• upload/download flow
• semantic exceptions
• schema cache
• structured pytest suite
• end-user usage examples

Still out of scope:

• automatic creation of connected items through throughConnectors
• public steps extensibility in updates and moves
• administrative operations for start form configuration

Vision

PipeBridge aims to be the standard semantic integration layer for Pipefy automation.

The current product direction is clear:

• keep the public facade small and coherent
• make common automation flows safer by default
• support extension without forcing forks
• keep documentation and examples strong enough for real adoption

Author

Rafael Mota Cavalcante

• GitHub: rmcavalcante7
• LinkedIn: rafael-cavalcante-dev-specialist
• E-mail: [e-mail]

License

MIT