Mycelium Search

BigQuery workflow: When unsure which table to query, call list_datasets → list_tables on the relevant dataset → get_table_schema on the specific table before writing SQL. Primary analytical dataset is mycelium_ai.

1. Financial Data → BigQuery (Xero)

Best first stop for any financial question (revenue, costs, P&L, balance sheet, budget).

Key tables

Schema notes

• entity = AU, UK, or US
• P&L group values: Income (AU/UK) or Revenue (US) for sales; Cost of Sales; Operating Expenses; Gross Profit; Net Profit
• Sales lines: line = 'Sales' and line = 'Sales - eCommerce'
• AU entity: amountUSD is often 0 — use amountAUD and convert if needed
• Row amounts also in amountInCurrency (native currency of entity)

Example query

SELECT entity, period, SUM(amountInCurrency) AS sales, currency
FROM `xero_processed_data.v_pnl_from_reports`
WHERE `group` IN ('Income', 'Revenue') AND line = 'Sales'
  AND period >= DATE_TRUNC(DATE_SUB(CURRENT_DATE(), INTERVAL 6 MONTH), MONTH)
GROUP BY entity, period, currency
ORDER BY period, entity

2. Inventory & Supply Chain → BigQuery (mycelium_ai dataset)

All cleaned Fishbowl tables live in the mycelium_ai dataset. Run list_tables('mycelium_ai') to see the full list.

mycelium_ai.t_inventory — Current inventory snapshot (yesterday only)

One row per inventory tag/lot per location. Key columns: entity, partNum, partName, partActiveFlag, productNum, productDescription, productPrice, locationGroupName, location, qtyOnHand, qtyCommitted, batchNum, expiryDate, bestBeforeDate, shelfLife_expDate, shelfLife_BestBeforeDate, costAUD, costGBP, costUSD, partCostInCurrency, currency, pack_size, assetaccount, tagId, partId, productId

• Available qty = qtyOnHand - qtyCommitted
• shelfLife_expDate buckets: (1) < 0 expired · (2) 0-99 · (3) 100-199 · (4) 200-299 · (5) 300-399 · (6) 400-499 · (7) 500-599 · (8) >= 600

mycelium_ai.t_inventory_history — Historical daily inventory (from 2022-12-05)

Same structure as t_inventory but with a date column. Use for trends and period comparisons.

mycelium_ai.t_sales — Sales order line items (all live orders, yesterday snapshot)

One row per sales order line item. Key columns: entity, soNum, soLineItem, soDateCreated, soDateIssued, soDateFulfilled, soFulfillmentDate, dateShipped, monthShipped, weekShipped, customerName, customerNumber, locationGroupName, soShipToName, soShipToCity, partNum, productNum, description, customerPartNum, qtyFulfilled, qtyToFulfill, unitsShipped, unitsToShip, soItemUnitPrice, soItemTotalPrice, totalWeight, soCustomerPo, soVendorPO, soSalesman, partActiveFlag, customerActiveFlag

mycelium_ai.t_purchase_orders — Purchase order headers

entity, ponumber, dateCreated, dateIssued, dateCompleted, postatus, vendorcontact, currency, totalTax, totalIncludesTax, carrier, username, paymentterms, shippingcountry

mycelium_ai.t_purchase_order_items — Purchase order line items

entity, ponumber, poLineItem, description, poitemtype, dateScheduledFulfillment, qtyFulfilled, qtypicked, qtytofulfill, uom, qtyfulfilledkg, unitcost, totalcost, currency Join to t_purchase_orders on ponumber.

mycelium_ai.t_transfer_orders — Transfer order headers

entity, xonumber, dateCreated, dateIssued, dateCompleted, xostatus, xotype, countryfrom, countryto, carrier, fromAttn, shiptoAttn

mycelium_ai.t_transfer_order_items — Transfer order line items

entity, xonumber, LineItem, description, xoitemtype, dateScheduledFulfillment, qtyFulfilled, qtypicked, qtytofulfill, uom, qtyfulfilledkg Join to t_transfer_orders on xonumber.

mycelium_ai.t_unit_price_history — Monthly part cost history (month-end snapshots)

entity, last_day_of_month, date, partNum, partName, avgCostInCurrency, currency, avgCostAUD

Raw Fishbowl datasources

Tables in fishbowl_datasources follow the pattern tablename (AU), tablename_uk, tablename_us — e.g. fishbowl_datasources.so, so_uk, so_us. Use mycelium_ai cleaned tables in preference.

3. CRM / Sales Pipeline → BigQuery (hubspot_datasources) or HubSpot live tools

Live lookups

Use HubSpot MCP tools (search_crm_objects, get_crm_objects) for real-time deal, contact, and company lookups. Object types: contacts, companies, deals.

Analytical queries: BigQuery hubspot_datasources

hubspot_datasources.c_deals hs_object_id, dealname, dealstage_name, pipeline_name, amount, amount_in_home_currency, deal_currency_code, closedate, end_deal_date, createdAt, hs_is_closed, hs_is_closed_won, hs_is_closed_lost, closed_won_reason, closed_lost_reason, hs_deal_stage_probability, hs_forecast_amount, hs_projected_amount, hs_acv, hs_arr, hs_mrr, hs_tcv, hs_days_to_close_raw, hs_is_stalled, hubspot_owner_name, hubspot_owner_assigneddate, hs_primary_associated_company, associations_contacts, associations_companies_count, associations_contacts_count, associations_emails_count

hubspot_datasources.c_contacts hs_object_id, firstname, lastname, email, phone, mobilephone, jobtitle, industry, lifecyclestage, lead_source, hs_buying_role, company, company_website, website, associatedcompanyid, address, city, state, country, linkedin, linkedin_profile, createdAt, lastmodifieddate

hubspot_datasources.c_companies hs_object_id, name, domain, website, industry, description, numberofemployees, hs_employee_range, no_outlets, annualrevenue, total_revenue, is_public, founded_year, address, city, state, country, timezone, phone, createdAt, updatedAt

hubspot_datasources.t_calls hs_object_id, hs_call_title, hs_call_summary, hs_body_preview, hs_call_direction, hs_call_status, hs_call_duration, hs_call_has_transcript, hs_call_has_voicemail, hs_call_primary_deal, hs_call_deal_stage_during_call, hubspot_owner_name, associations_deals, createdAt

hubspot_datasources.t_emails hs_object_id, hs_email_subject, hs_email_text, hs_body_preview, hs_email_sender_email, hs_email_direction, hs_email_status, hs_email_open_count, hs_email_reply_count, hs_email_click_count, hs_email_associated_contact_id, hubspot_owner_name, associations_deals, createdAt

hubspot_datasources.t_meetings hs_object_id, hs_meeting_title, hs_meeting_body, hs_meeting_start_time, hs_internal_meeting_notes, hs_body_preview, hubspot_owner_name, associations_deals, createdAt

hubspot_datasources.t_notes hs_object_id, hs_note_body, hs_body_preview, hubspot_owner_name, associations_deals, createdAt

4. Policies, Procedures & Reference Docs → Confluence

Site: https://fablefood.atlassian.net/wiki · Cloud ID: 9aaea3b1-0b23-49a9-8f86-a7f2f02ca748

How to search and cite

1. Search first: use search_pages with CQL before browsing
2. Get page content: use get_page with the page ID
3. Check attachments: use get_page_attachments then get_attachment_content — contracts, price lists, specs and SOPs are often attachments not inline text
4. Cite with URL: https://fablefood.atlassian.net/wiki/spaces/{KEY}/pages/{PAGE_ID}
5. Explore hierarchy: use get_page_children for spaces with nested structures

CQL examples

space = "FIN" AND title ~ "payroll"
space = "CUS" AND title ~ "HelloFresh"
text ~ "FB003" AND space = "FQS"
text ~ "Sysco" AND space = "LEG"
space = "MAN" AND title ~ "Hoshay"

Spaces

5. Team Communication & Real-Time Updates → Slack

Workspace: fungideas.slack.com

Use search_messages with keywords. Use read_thread for full reply context. Cite channel name, sender, and date.

Sales & Pipeline

Customer-specific channels

Operations & Logistics

Finance

Product, Quality & R&D

Marketing & PR

Legal

Team & Culture

6. Meeting Notes & Transcripts → Fireflies

Admin user: Ji Davis ([email protected]). Always cite: meeting title, date, and Fireflies transcript ID.

7. Working Files & Reference Documents → Google Drive

Google Drive holds live working files, trackers, templates and reference documents across 6 shared drives. All files returned by the connector are broadly accessible (shared drive members, domain-wide, or anyone-with-link) — restricted personal files are automatically filtered out.

How to search and cite

1. Search by keyword: search_files with a name or content query
2. Browse a drive: list_shared_drives to get drive IDs → list_files with the drive's folder ID
3. Read content: read_file exports Google Docs as plain text, Sheets as CSV, Slides as plain text (10 MB limit)
4. Cite with link: use the webViewLink returned in every result

Shared drives and what lives there

What Drive is best for (vs Confluence)

• Drive → live working files, active trackers, models being iterated on, large media assets, presentations
• Confluence → finalised policies, SOPs, reference documentation, structured knowledge base articles

Example search queries

# Find sales target or attribution trackers
search_files(query="name contains 'target' and mimeType = 'application/vnd.google-apps.spreadsheet'")

# Find customer pitch decks in the Sales drive
search_files(query="name contains 'Fable' and mimeType = 'application/vnd.google-apps.presentation'", shared_drive_id="0ADuDVnKjPGVnUk9PVA")

# Find budget or forecast working files
search_files(query="name contains 'budget' or name contains 'forecast'", include_shared_drives=true)

# Browse the Finance drive root
list_files(folder_id="0AKjO6l0ehgnbUk9PVA")

8. External Email → Gmail

• search_emails: filter by from:, to:, subject:, date range
• get_email: read a full email thread
• Defaults to external-only (filters out @fablefood.co)
• Use user_email parameter to read any team member's inbox

Best for: customer order confirmations, supplier negotiations, inbound enquiries, distributor/retailer threads.

Quick Decision Guide

Citing sources

• BigQuery: full table name (e.g. `mycelium_ai.t_inventory`) + query used
• Confluence page: title + URL (https://fablefood.atlassian.net/wiki/spaces/{KEY}/pages/{ID})
• Confluence attachment: filename + parent page URL
• Slack: channel name, sender, approximate date
• Fireflies: meeting title, date, transcript ID
• Gmail: sender, recipient, subject, date
• Google Drive: file name + webViewLink (e.g. https://docs.google.com/spreadsheets/d/{ID}/edit)

Last updated: March 2026. Maintained by Ji Davis ([email protected]). Update when new spaces, channels, datasets, or connectors are added.