JSON format expected from general peer [GEN -> DEP]

[centauri1219]

The depth peer will require the following JSON format from the general peer at each depth iteration in the case of querying.

{
  "source": "genpeer",
  "source_id":123
  "embed": [0.45, 0.78, ...],    
  "prev_depth": 2,
  "query_type": "search",           
  "threshold": 0.75,
  "results_count": 5                     
}

Key	Type	Example	Description
source	string	"genpeer"	Name of Source
source_id	int	44	Unique ID of the source
embed	[]float64	[0.45, 0.78, ...]	Embedding vector representing the query
prev_depth	int	2	Keeps track of current depth
query_type	string	"index"	To differentiate between indexing and querying
threshold	float64	0.60	For similarity analysis
results_count	int	3	No. of results to be fetched (for D4 only)

In the case of indexing we also need file metadata.

{
  "source": "genpeer",
  "source_id": 34392
  "embed": [0.23, 0.91, ...],               
  "depth": 2,                              
  "file_metadata": {
    "name": "project_report_final.pdf",
    "created_at": "2025-07-01T09:15:00Z",
    "updated_at": "2025-07-05T14:45:00Z"
  },
  "query_type": "index",                    // explicitly mention the purpose
  "threshold": 0.8                          // similarity threshold to accept
}

Key	Type	Example	Description
source	string	"genpeer"	Name of Source
source_id	int	44	Unique ID of the source
embed	[]float64	[0.45, 0.78, ...]	Embedding vector representing the new file
depth	int	2	Keeps track of current depth
file_metadata	obj	Refer below table	Keep track of the file metadata to be stored
query_type	string	"index"	To differentiate between indexing and querying
threshold	float64	0.60	For similarity analysis
results_count	int	3	No. of results to be fetched (for D4 only)

Key	Type	Example	Description
name	string	"document.pdf"	Name of the file
created_at	string	"2025-07-05T12:00:00Z"	ISO timestamp of file creation
updated_at	string	"2025-07-05T13:00:00Z"	ISO timestamp of last update

@YugDalwadi @SiddhantGahankari @Abhinav-Kumar2 @IshatV412 Check out the fields and do suggest changes on additional or different return types that may be required.

[xyphoes0727]

1. Won't the Depth 4 peer return results_count?
2. What's the use of prev_depth?

[IshatV412]

1. Yes D4 peer will return results_count. We forgot to add the comment.
2. prev_depth is just for easier logging of data being transmitted and easier error handling

[centauri1219]

Adding on to 2. prev_depth is just for keeping the updated depth information at each relay back so that the gen_peer will be able to understand that if the depth is 4, it can stop it's peer contacting process and directly contact the owner peers based on the results presented.

[IshatV412]

we also need a target peerID key in the request. the target peerID is required in the routing logic. the target peerID refers to the peer with which you want to communicate eventually. For eg-: if I want to communicate with a D1 peer, i need the peerID of that peer in the query as well.