Skip to content

Commit ed8bc1b

Browse files
maparentmaparent
committed
ENG-1822 modify upsert
1 parent 13728b7 commit ed8bc1b

7 files changed

Lines changed: 629 additions & 188 deletions

File tree

‎packages/database/doc/upsert.md‎

Lines changed: 236 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,236 @@
1+
# A few examples of using upsert_authors / upsert_document / upsert_content / upsert_concept etc.
2+
3+
In general, for external references such as `author_id`, you can either:
4+
5+
1. use the database id as is: `author_id`
6+
2. Use the platform id, with the name transformed such as `author_local_id`
7+
3. Put the data inline as a subobject, with the name transformed such as `author_inline`
8+
9+
Note that embeddings are always inline.
10+
11+
Here are complete examples:
12+
13+
```typescript
14+
import type {
15+
LocalAccountDataInput,
16+
LocalDocumentDataInput,
17+
LocalContentDataInput,
18+
LocalConceptDataInput,
19+
} from "@repo/database/inputTypes";
20+
import type { Json } from "@repo/database/dbTypes";
21+
import type { DGSupabaseClient } from "@repo/database/lib/client";
22+
23+
const demoCode = async (client: DGSupabaseClient) => {
24+
const accounts: LocalAccountDataInput[] = [
25+
{
26+
account_local_id: "sR22zZ470dNPkIf9PpjQXXdTBjG2",
27+
name: "maparent",
28+
},
29+
];
30+
31+
const docs: LocalDocumentDataInput[] = [
32+
{
33+
source_local_id: "page1_uid",
34+
created: "2000/01/01",
35+
last_modified: "2001/01/02",
36+
author_local_id: "sR22zZ470dNPkIf9PpjQXXdTBjG2",
37+
},
38+
];
39+
40+
const contents: LocalContentDataInput[] = [
41+
{
42+
author_local_id: "sR22zZ470dNPkIf9PpjQXXdTBjG2",
43+
author_inline: {
44+
account_local_id: "sR22zZ470dNPkIf9PpjQXXdTBjG2",
45+
name: "maparent",
46+
},
47+
document_inline: {
48+
source_local_id: "page1_uid",
49+
created: "2000/01/01",
50+
last_modified: "2001/01/02",
51+
author_local_id: "sR22zZ470dNPkIf9PpjQXXdTBjG2",
52+
},
53+
source_local_id: "a_roam_uid",
54+
scale: "document",
55+
created: "2000/01/01",
56+
last_modified: "2001/01/02",
57+
text: "[[CLM]] a claim",
58+
},
59+
{
60+
author_local_id: "sR22zZ470dNPkIf9PpjQXXdTBjG2",
61+
document_local_id: "page1_uid",
62+
source_local_id: "a_roam_uid2",
63+
scale: "section",
64+
created: "2000/01/02",
65+
last_modified: "2001/01/03",
66+
part_of_local_id: "a_roam_uid",
67+
text: "Some subsection",
68+
},
69+
{
70+
author_local_id: "sR22zZ470dNPkIf9PpjQXXdTBjG2",
71+
document_inline: docs[0],
72+
source_local_id: "a_roam_uid3",
73+
scale: "paragraph",
74+
created: "2000/01/02",
75+
last_modified: "2001/01/03",
76+
part_of_local_id: "a_roam_uid2",
77+
text: "Some paragraph",
78+
embedding_inline: {
79+
model: "openai_text_embedding_3_small_1536",
80+
vector: [0], // assume that the vector has the requisite length
81+
},
82+
},
83+
];
84+
85+
const concepts: LocalConceptDataInput[] = [
86+
{
87+
author_local_id: "sR22zZ470dNPkIf9PpjQXXdTBjG2",
88+
source_local_id: "a_roam_uid3",
89+
created: "2000/01/01",
90+
last_modified: "2001/01/02",
91+
name: "Some subsubtext",
92+
schema_represented_by_local_id: "known_claim_schema_local_id",
93+
},
94+
];
95+
96+
// Base scenario: insert references in order of dependencies
97+
// 1. upsert accounts.
98+
{
99+
const { data, error } = await client.rpc("upsert_accounts_in_space", {
100+
space_id_: 12,
101+
accounts,
102+
});
103+
if (error) console.error(error);
104+
console.log(data);
105+
}
106+
107+
// 2. upsert documents.
108+
{
109+
const { data, error } = await client.rpc("upsert_documents", {
110+
v_space_id: 12,
111+
data: docs as Json,
112+
});
113+
if (error) console.error(error);
114+
console.log(data);
115+
}
116+
117+
// 3. content
118+
{
119+
const { data, error } = await client.rpc("upsert_content", {
120+
v_space_id: 12,
121+
data: contents as Json,
122+
v_creator_id: 63,
123+
});
124+
if (error) console.error(error);
125+
console.log(data);
126+
}
127+
128+
// 4. upsert concept
129+
{
130+
const { data, error } = await client.rpc("upsert_concepts", {
131+
v_space_id: 12,
132+
data: concepts as Json,
133+
});
134+
if (error) console.error(error);
135+
console.log(data);
136+
}
137+
138+
// Variant: if all content is known to also be a document, you can upsert both as if it were a single object
139+
const page_contents: LocalContentDataInput[] = [
140+
{
141+
author_local_id: "sR22zZ470dNPkIf9PpjQXXdTBjG2",
142+
source_local_id: "a_page_uid2",
143+
created: "2000/01/02",
144+
last_modified: "2001/01/03",
145+
text: "Some other page",
146+
},
147+
];
148+
149+
{
150+
const { data, error } = await client.rpc("upsert_content", {
151+
v_space_id: 12,
152+
data: page_contents as Json,
153+
v_creator_id: 63,
154+
content_as_document: true,
155+
});
156+
if (error) console.error(error);
157+
console.log(data);
158+
}
159+
160+
// Nesting: you can nest information. Nested types can inherit information from the type above.
161+
// Eg 1: upserting content with nested documents and authors
162+
const contentWithNestedDoc: LocalContentDataInput[] = [
163+
{
164+
author_local_id: "sR22zZ470dNPkIf9PpjQXXdTBjG2",
165+
source_local_id: "a_page_uid3",
166+
created: "2000/01/02",
167+
scale: "document",
168+
last_modified: "2001/01/03",
169+
text: "Yet another page",
170+
author_inline: {
171+
account_local_id: "sR22zZ470dNPkIf9PpjQXXdTBjG2",
172+
name: "maparent",
173+
},
174+
document_inline: {
175+
source_local_id: "a_page_uid3",
176+
// note that created, last_modified and author_inline will be inherited
177+
},
178+
},
179+
];
180+
{
181+
const { data, error } = await client.rpc("upsert_content", {
182+
v_space_id: 12,
183+
data: contentWithNestedDoc as Json,
184+
v_creator_id: 63,
185+
});
186+
if (error) console.error(error);
187+
console.log(data);
188+
}
189+
190+
// eg 2: Nesting multiple contents in a concept
191+
const conceptsWithNestedContent: LocalConceptDataInput[] = [
192+
{
193+
author_local_id: "sR22zZ470dNPkIf9PpjQXXdTBjG2",
194+
source_local_id: "a_roam_uid4",
195+
created: "2000/01/01",
196+
last_modified: "2001/01/02",
197+
name: "[[CLM]] another claim",
198+
schema_represented_by_local_id: "known_claim_schema_local_id",
199+
author_inline: {
200+
account_local_id: "sR22zZ470dNPkIf9PpjQXXdTBjG2",
201+
name: "maparent",
202+
},
203+
document_inline: {
204+
source_local_id: "a_page_uid3",
205+
// note that created, last_modified and author_inline will be inherited
206+
},
207+
contents_inline: [
208+
{
209+
text: "The claim page contents",
210+
variant: "full",
211+
// here, source_local_id, document_inline, author_inline, created, last_modified are inherited
212+
},
213+
{
214+
text: "[[CLM]] another claim",
215+
variant: "direct",
216+
embedding_inline: {
217+
model: "openai_text_embedding_3_small_1536",
218+
vector: [0], // assume that the vector has the requisite length
219+
// again source_local_id etc. are inherited
220+
},
221+
// same variables are intherited
222+
},
223+
],
224+
},
225+
];
226+
227+
{
228+
const { data, error } = await client.rpc("upsert_concepts", {
229+
v_space_id: 12,
230+
data: conceptsWithNestedContent as Json,
231+
});
232+
if (error) console.error(error);
233+
console.log(data);
234+
}
235+
};
236+
```

‎packages/database/doc/upsert_content.md‎

Lines changed: 0 additions & 69 deletions
This file was deleted.

0 commit comments

Comments
 (0)