Outbound integrations

Get findings out — to ticketing, chat, email, and BI/warehouse — without coupling to any one vendor. Everything here operates on the findings layer and is read-only toward the BAS: it pushes results to people and systems, never to a controller.

flowchart LR
  findings["Findings"]
  ticket["finding_to_ticket"]
  dispatch["dispatch_findings (severity filter, dedupe)"]
  fmt["format_for"]
  webhook["webhook / CMMS"]
  slack["slack"]
  teams["teams"]
  export["export_findings"]
  frame["findings_to_frame"]
  files["csv / json / parquet"]
  findings --> ticket
  ticket --> dispatch
  dispatch --> fmt
  fmt --> webhook
  fmt --> slack
  fmt --> teams
  findings --> export
  export --> frame
  frame --> files

Findings render to tickets, then dispatch to chat/CMMS channels or export to a tabular warehouse frame.

camber.integrate provides three layers:

1. Ticket render (tickets.py) — finding_to_ticket / findings_to_tickets map a Finding to a neutral, JSON-serializable CMMS work-order dict (severity→priority, stable fingerprint for dedupe). webhook_transport(url) and collect_transport() (in-memory, for tests/dry-runs).
2. Channels + dispatch (notify.py).
3. Tabular export (export.py).

Dispatch findings to a channel

dispatch_findings(findings, transport, *, channel, min_severity, site, seen, dedupe, dry_run) renders, filters by severity, dedupes, formats per channel, and sends.

from camber.integrate import dispatch_findings, webhook_transport, email_transport

# Slack incoming webhook, only warn+ , deduped across runs
seen = load_seen()  # a set of fingerprints you persist
dispatch_findings(
    findings,
    webhook_transport(SLACK_URL),
    channel="slack",
    min_severity="warn",
    site="HQ",
    seen=seen,
)
save_seen(seen)

# Email via SMTP (raw ticket -> subject/body)
dispatch_findings(
    findings,
    email_transport(
        "smtp.example.com",
        sender="camber@x.com",
        recipients=["ops@x.com"],
        username="u",
        password="p",
    ),
    channel="webhook",
    min_severity="fault",
)

# CMMS / generic webhook: channel="webhook" posts the raw ticket dict
dispatch_findings(findings, webhook_transport(CMMS_URL), channel="webhook")

Channels — format_for(channel, ticket)

• webhook — the raw ticket dict (for a generic JSON webhook / CMMS adapter / email).
• slack — slack_payload: {"text": …} with a priority emoji.
• teams — teams_payload: a Teams MessageCard with color + facts.

Option flags — dispatch_findings

flag	default	effect
channel	"webhook"	payload shape (webhook / slack / teams)
min_severity	"warn"	drop findings below this (ok/info/warn/fault)
seen	None	a fingerprint set you persist; with dedupe, only new findings send
dedupe	True	skip findings whose fingerprint is already in seen
dry_run	False	format but don't call the transport (returns what would send)
site	""	stamped into the ticket / fingerprint

email_transport(host, *, port, sender, recipients, use_tls, username, password, _smtp_factory) is a transport (subject = ticket title, body = ticket body); _smtp_factory is injectable for tests.

Export for BI / warehouse

from camber.integrate import export_findings, findings_to_frame

df = findings_to_frame(findings, site="HQ")  # one row per finding, metrics flattened
export_findings(findings, "findings.parquet", site="HQ")  # csv / json / parquet by extension

Option flags — export_findings / findings_to_frame

flag	default	effect
format	by extension	csv / json (records) / parquet
flatten_metrics	True	metrics → metric_<key> scalar columns
columns	all	restrict/order the output columns
site	""	site column + fingerprint stamp