JSONMemoryStore¶
JSONL-file-backed episodic memory store.
JSONMemoryStore
¶
Bases: BaseMemoryStore
Episodic memory store backed by a JSONL file on disk.
Each episode is stored as one JSON object per line (JSONL format).
New episodes are appended to the file — no full rewrite on each write.
The full file is read into memory on construction and on read_recent.
Attributes:
| Name | Type | Description |
|---|---|---|
path |
Path
|
Full path to the backing JSONL file ( |
Source code in src/llm_agents_from_scratch/memory/json_store.py
10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 | |
__init__
¶
Initialize a JSONMemoryStore.
Loads any existing episodes from dir / filename on construction.
The file is created on the first write call if it does not yet
exist.
Parameters:
| Name | Type | Description | Default |
|---|---|---|---|
dir
|
Path
|
Directory in which the backing JSONL file is stored. The caller decides the location; no default is imposed by the library. |
required |
filename
|
str
|
Name of the JSONL file within |
'episodes.jsonl'
|
Source code in src/llm_agents_from_scratch/memory/json_store.py
write
async
¶
Persist an episode to the store.
Appends one JSON line to the backing file. Does not rewrite existing
content. embedded_text is accepted for interface compatibility
but ignored — this store does not embed episodes.
Parameters:
| Name | Type | Description | Default |
|---|---|---|---|
episode
|
Episode
|
The completed episode to store. |
required |
embedded_text
|
str | None
|
Ignored. |
None
|
Source code in src/llm_agents_from_scratch/memory/json_store.py
read_recent
async
¶
Return the N most recently recorded episodes.
Reads all episodes from the in-memory list. A production implementation would tail the file to avoid reading it in full.
Parameters:
| Name | Type | Description | Default |
|---|---|---|---|
n
|
int
|
Maximum number of episodes to return. |
required |
Returns:
| Type | Description |
|---|---|
list[Episode]
|
list[Episode]: Episodes ordered from most recent to oldest. |
Source code in src/llm_agents_from_scratch/memory/json_store.py
count
async
¶
Return the total number of episodes in the store.
Returns:
| Name | Type | Description |
|---|---|---|
int |
int
|
Episode count. |
summary
async
¶
Return a human-readable summary of the store contents.
Includes the backing file path, total episode count, and the instruction and timestamp of the newest and oldest episodes.
Returns:
| Name | Type | Description |
|---|---|---|
str |
str
|
Multi-line summary of the store. |
Source code in src/llm_agents_from_scratch/memory/json_store.py
search
async
¶
Not implemented — similarity search is deferred to vector store.
Parameters:
| Name | Type | Description | Default |
|---|---|---|---|
query
|
str
|
The search query. |
required |
k
|
int
|
Maximum number of episodes to return. |
required |
**kwargs
|
Any
|
Ignored. |
{}
|
Raises:
| Type | Description |
|---|---|
NotImplementedError
|
Always. Use a vector-backed store for similarity search. |