field_configs.yaml

# field_configs.yaml - PTAB MCP Field Configuration
# Progressive disclosure: minimal → balanced → complete for each data type
# Customize field sets by uncommenting fields you want (remove the # symbol)
#
# ⚠️ IMPORTANT: After editing this file, you MUST restart the MCP server for changes to take effect.
# In Claude Desktop: Close Claude completely and reopen it. The new field configuration will be loaded on startup.
# The search tools will then use your updated field sets.
#
# USAGE:
# 1. Find the field set you want to customize (e.g., trials_minimal)
# 2. Uncomment any field by removing the # at the start of the line
# 3. Save this file
# 4. Restart Claude Desktop completely
# 5. The new fields will be included in search results

version: "1.0"
description: "PTAB MCP Field Configuration - Progressive disclosure"

predefined_sets:
  # ============================================================================
  # TRIALS (IPR, PGR, CBM Proceedings)
  # ============================================================================
  trials_minimal:
    description: "Ultra-minimal fields for trial discovery (95-99% context reduction)"
    fields:
      # === MOST USED FIELDS ===
      - trialNumber                                        # Trial number (IPR2024-00123)
      - trialMetaData.accordedFilingDate                   # Filing date
      - trialMetaData.trialTypeCode                        # IPR, PGR, CBM
      - trialMetaData.trialStatusCategory                  # Status
      - trialMetaData.institutionDecisionDate              # Institution decision date
      - trialMetaData.terminationDate                      # Final decision date
      - regularPetitionerData.realPartyInInterestName      # Petitioner name (CORRECTED)
      - regularPetitionerData.counselName                  # Petitioner counsel
      - patentOwnerData.patentOwnerName                    # Patent owner name
      - patentOwnerData.patentNumber                       # Patent number (CORRECTED)
      - patentOwnerData.grantDate                          # Patent grant date
      - patentOwnerData.applicationNumberText              # Application number

      # === ADDITIONAL TRIAL METADATA (uncomment to add) ===
      # - trialMetaData.petitionFilingDate       # Petition filing date
      # - trialMetaData.trialLastModifiedDate    # Last modified date
      # - trialMetaData.trialLastModifiedDateTime # Last modified datetime
      # - trialMetaData.fileDownloadURI          # File download URI

      # === PARTY INFORMATION (uncomment to add) ===
      # - regularPetitionerData.realPartyInInterestName  # Real party in interest (already in minimal)
      # - regularPetitionerData.counselName              # Petitioner counsel (already in minimal)
      # - patentOwnerData.realPartyInInterestName        # Patent owner real party in interest
      # - patentOwnerData.counselName                    # Patent owner counsel

      # === PATENT/APPLICATION INFORMATION (uncomment to add) ===
      # - patentOwnerData.inventorName           # Inventor name
      # - patentOwnerData.technologyCenterNumber # Technology center
      # - patentOwnerData.groupArtUnitNumber     # Art unit

      # NOTE: The API does not currently provide decision data, claims data, or detailed
      # classification data in the trials endpoint. Use ptab_get_trial_documents() to
      # access full decision documents.

      # === DOCUMENTS (uncomment to add) ===
      # WARNING: documentBag can cause massive token usage (similar to PFW's documentBag)
      # Use ptab_get_trial_documents() tool for targeted document access instead
      # - documentBag  # ⚠️ WARNING: Can cause 100x token increase - use ptab_get_trial_documents instead

  trials_balanced:
    description: "Comprehensive trial analysis (90-95% context reduction)"
    fields:
      # All fields via wildcard patterns
      - trialNumber                  # Trial number
      - lastModifiedDateTime         # Last modified datetime
      - trialMetaData.*              # All trial metadata
      - regularPetitionerData.*      # All petitioner data (CORRECTED)
      - patentOwnerData.*            # All patent owner and patent data (CORRECTED)

  trials_complete:
    description: "Complete trial data (80-90% context reduction)"
    fields:
      - "*"                          # All fields

  # ============================================================================
  # APPEALS (Ex Parte Appeals to PTAB)
  # ============================================================================
  appeals_minimal:
    description: "Ultra-minimal fields for appeal discovery (95-99% context reduction)"
    fields:
      # === MOST USED FIELDS ===
      - appealNumber                             # Appeal number
      - applicationNumber                        # Application number
      - documentData.documentFilingDate          # Filing date
      - documentData.decisionDate                # Decision date
      - appellantData.technologyCenterNumber     # Technology center
      - appellantData.groupArtUnitNumber         # Art unit
      - documentData.decisionTypeCodeDescription # Decision type
      - documentData.decisionOutcome             # Outcome (affirmed, reversed, etc.)
      - appellantData.appellantName              # Appellant name

      # === ADDITIONAL APPEAL METADATA (uncomment to add) ===
      # - documentData.documentIdentifier        # Document ID
      # - documentData.documentCode              # Document code
      # - documentData.documentDescription       # Document description
      # - documentData.boardDecisionIndicator    # Board decision indicator (Y/N)

      # === DATES (uncomment to add) ===
      # - documentData.mailDate                  # Mail date
      # - documentData.entryDate                 # Entry date
      # - documentData.publicationDate           # Publication date

      # === APPELLANT INFORMATION (uncomment to add) ===
      # - appellantData.appellantAddress         # Appellant address
      # - appellantData.appellantCounsel         # Appellant attorney/counsel
      # - appellantData.appellantCounselPhone    # Counsel phone number
      # - appellantData.appellantType            # Appellant type

      # === EXAMINER INFORMATION (uncomment to add) ===
      # - examinerData.primaryExaminerName       # Primary examiner
      # - examinerData.assistantExaminerName     # Assistant examiner
      # - examinerData.examinerArtUnit           # Examiner art unit

      # === DECISION DETAILS (uncomment to add) ===
      # - decisionData.claimsAppealed            # Claims appealed
      # - decisionData.claimsAffirmed            # Claims affirmed
      # - decisionData.claimsReversed            # Claims reversed
      # - decisionData.claimsReversedInPart      # Claims reversed in part
      # - decisionData.decisionSummary           # Decision summary

      # === PATENT/APPLICATION DATA (uncomment to add) ===
      # - applicationData.inventionTitle         # Invention title
      # - applicationData.filingDate             # Application filing date
      # - applicationData.patentNumber           # Patent number (if granted)
      # - applicationData.uspcClassification     # US Patent Classification
      # - applicationData.cpcClassification      # Cooperative Patent Classification

      # === DOCUMENTS (uncomment to add) ===
      # Use ptab_get_appeal_documents() tool for targeted document access
      # - documentBag  # ⚠️ WARNING: Can cause large token usage - use ptab_get_appeal_documents instead

  appeals_balanced:
    description: "Comprehensive appeal analysis (90-95% context reduction)"
    fields:
      # All minimal fields via wildcard patterns
      - appealNumber
      - applicationNumber
      - documentData.*               # All document/decision data
      - appellantData.*              # All appellant data
      - examinerData.*               # All examiner data
      - decisionData.*               # All decision details

  appeals_complete:
    description: "Complete appeal data (80-90% context reduction)"
    fields:
      - "*"                          # All fields

  # ============================================================================
  # INTERFERENCES (Interference Proceedings)
  # ============================================================================
  interferences_minimal:
    description: "Ultra-minimal fields for interference discovery (95-99% context reduction)"
    fields:
      # === MOST USED FIELDS ===
      - interferenceNumber                       # Interference number
      - documentData.documentFilingDate          # Filing date
      - documentData.decisionDate                # Decision date
      - documentData.decisionType                # Decision type
      - partyData.seniorParty                    # Senior party
      - partyData.juniorParty                    # Junior party

      # === ADDITIONAL INTERFERENCE METADATA (uncomment to add) ===
      # - documentData.documentIdentifier        # Document ID
      # - documentData.documentCode              # Document code
      # - documentData.documentDescription       # Document description
      # - documentData.decisionOutcome           # Decision outcome

      # === DATES (uncomment to add) ===
      # - documentData.mailDate                  # Mail date
      # - documentData.entryDate                 # Entry date
      # - documentData.publicationDate           # Publication date
      # - documentData.judgmentDate              # Judgment date

      # === PARTY INFORMATION (uncomment to add) ===
      # - partyData.seniorPartyAddress           # Senior party address
      # - partyData.seniorPartyCounsel           # Senior party counsel
      # - partyData.juniorPartyAddress           # Junior party address
      # - partyData.juniorPartyCounsel           # Junior party counsel

      # === PATENT/APPLICATION DATA (uncomment to add) ===
      # - partyData.seniorPartyPatentNumber      # Senior party patent number
      # - partyData.seniorPartyApplicationNumber # Senior party application number
      # - partyData.juniorPartyPatentNumber      # Junior party patent number
      # - partyData.juniorPartyApplicationNumber # Junior party application number
      # - patentData.inventionTitle              # Invention title

      # === DECISION DETAILS (uncomment to add) ===
      # - decisionData.priority                  # Priority determination
      # - decisionData.claimsInInterference      # Claims in interference
      # - decisionData.decisionSummary           # Decision summary

      # === DOCUMENTS (uncomment to add) ===
      # Use ptab_get_interference_documents() tool for targeted document access
      # - documentBag  # ⚠️ WARNING: Can cause large token usage - use ptab_get_interference_documents instead

  interferences_balanced:
    description: "Comprehensive interference analysis (90-95% context reduction)"
    fields:
      # All minimal fields via wildcard patterns
      - interferenceNumber
      - documentData.*               # All document/decision data
      - partyData.*                  # All party data
      - decisionData.*               # All decision details

  interferences_complete:
    description: "Complete interference data (80-90% context reduction)"
    fields:
      - "*"                          # All fields

# ============================================================================
# Context Management Settings
# ============================================================================
context_settings:
  minimal_reduction_percentage: 95
  balanced_reduction_percentage: 85
  complete_reduction_percentage: 80
  max_field_count_minimal: 15
  max_field_count_balanced: 50

# ============================================================================
# Usage Instructions
# ============================================================================
# 1. TRIAL SEARCHES:
#    - Edit 'trials_minimal' for ultra-fast discovery (50+ results)
#    - Edit 'trials_balanced' for detailed analysis (10-20 results)
#    - Use 'trials_complete' for comprehensive data
#
# 2. APPEAL SEARCHES:
#    - Edit 'appeals_minimal' for ultra-fast discovery
#    - Edit 'appeals_balanced' for detailed analysis
#    - Use 'appeals_complete' for comprehensive data
#
# 3. INTERFERENCE SEARCHES:
#    - Edit 'interferences_minimal' for ultra-fast discovery
#    - Edit 'interferences_balanced' for detailed analysis
#    - Use 'interferences_complete' for comprehensive data
#
# 4. FIELD CUSTOMIZATION:
#    - Add/remove fields from any set
#    - Use wildcard patterns (e.g., "trialMetaData.*" for all trial metadata)
#    - Restart MCP server after editing
#
# 5. WILDCARD PATTERNS:
#    - "trialMetaData.*" - All trial metadata fields
#    - "petitionerData.*" - All petitioner data fields
#    - "respondentData.*" - All respondent/patent data fields
#    - "*" - All available fields (complete tier)
#
# 6. CONTEXT REDUCTION:
#    - Minimal tier: 95-99% reduction (10-15 fields)
#    - Balanced tier: 85-95% reduction (30-50 fields)
#    - Complete tier: 80-90% reduction (all fields with API-side filtering)