Skip to content

Commit 1e03837

Browse files
valhubervalhuber
committed
import doc
1 parent 1e0d99a commit 1e03837

File tree

3 files changed

+133
-9
lines changed

3 files changed

+133
-9
lines changed

‎docs/IDE-Fixup-Attrs.md

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -7,7 +7,7 @@
77

88
1. Collects the latest model, rules, and test data from the project.
99
2. Calls ChatGPT (or similar) to resolve missing columns or data in the project.
10-
3. Saves the fixup request/response under a 'fixup' folder.
10+
3. Saves the fixup request/response under a `docs/fixup` folder.
1111
4. You then use this to create a new project
1212

1313
## Using Fixup

‎docs/IDE-Import-WebGenAI.md

Lines changed: 122 additions & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -1,7 +1,127 @@
11
!!! pied-piper ":bulb: TL;DR - Import WebGenAI Project"
22

3-
You can import WegGenAI projects, merging their data models and rules into a dev project.
3+
You can import WebGenAI projects, merging their data models and rules into a dev project.
44

5-
Under construction, coming soon.
5+
`als genai-utils --import-genai --using=<WebGenAI-Project>`
66

77
&nbsp;
8+
9+
# Coordinating Parallel Dev Streams (Multi-Team Development)
10+
11+
This is the Diego Lo Giudice challenge: enable ongoing parallel development with both the LOB and Dev *teams.* It's enabled by declarative technology, where the integration is done with software, not manual effort.
12+
13+
It works like this:
14+
15+
1. The project begins with the **Kickstart:** several iterations to get *solid requirements*.
16+
* Uses Natural Language and Declarative Rules
17+
18+
2. The project is **exported:** the Dev Team begin work on the Custom UI, Enterprise Integration, etc
19+
* This is *not* the end of declarative: logic is either or both of Natural Language, and Python DSL
20+
21+
3. As depicted in the **divergent arcs:** there are now 2 versions of the project. Both can, in parallel, introduce new rules and attributes. This is not just "Team Development", it is ***TeamS Development***.
22+
* **LOB Parallel:** (WG_Team): new rules & attributes
23+
* Observe: WG is not a day-1-only pilot...
24+
* They continue to use the Natural Language Web-Based interface (WebGenai)
25+
* **Dev_Team:** their own new rules & attributes, using either genai Natural Language and/or alembic...
26+
* In addition, the server team also uses Python (and) and libs as required, e.g. for enterprise integration.
27+
* They do this in their favorite IDE, with tools such as GitHub etc.
28+
* Observe the logic remains at a high lev
29+
* The UI team can use familiar tools for Custom UIs. These
30+
* leverage the API (ready day 1, so they are not blocked on API Dev), and
31+
* are dramatically simplified by automated backend rule-based logic
32+
33+
4. **Export 2:** illustrates that the LOB WG_Team can export their project. The dev team can import it using the API Logic Server CLI.
34+
35+
* This automatically integrates rules and attributes from both projects, **updating the dev project with a new database and models.**
36+
* This is virtually impossible with procedural code, because developers must manually assess the execution dependencies and order the logic properly. It time-consuming, complex, and error-prone - just like post deployment maintenance.
37+
* The integration is automatic and "safe" because logic is expressed in a declarative rules/models for which *ordering is automatic.*
38+
* It also rebuilds the test data, per your rules (e.g. sum/count values)
39+
* The process supports multiple exports.
40+
41+
<br/>
42+
43+
![genai-process](images/sample-ai/genai/genai-process.png)
44+
45+
<br/>
46+
47+
# Exploring Import
48+
49+
<br/>
50+
51+
## Setup: Manager pre-installs Import Sample
52+
53+
When [you create the manager (**strongly recommended**)](https://apilogicserver.github.io/Docs/Manager/), the system installs 3 sample projects you can use to explore import.
54+
55+
1. **Base Project** is GenAI_no_logic. No rule-based attributes. See `system/genai/examples/genai_demo/wg_dev_merge/base_genai_demo_no_logic`. It's not really used, just provided as a reference.
56+
57+
2. **Dev Project** was created with export-1, and has added rules for `carbon_neutral`. It is ready for export-2. See `system/genai/examples/genai_demo/wg_dev_merge/dev_demo_no_logic_fixed`
58+
59+
3. **WG project** has continued from export-1 to add our standard customer.balance rules. It is ready for export-2. See `system/genai/examples/genai_demo/wg_dev_merge/wg_demo_no_logic_fixed`.
60+
61+
* It has an `docs/export/export.json`, which describes the data model and rules from the WG project. This is used for import.
62+
63+
The naming convention is that these started with no rules, had rules added, and were **"fixed"** by Genai-Logic to update the data model.
64+
65+
<br/>
66+
67+
## Usage
68+
69+
Imports are performed from with the dev project, using the `import-genai` CLI command:
70+
71+
```bash
72+
cd system/genai/examples/genai_demo/wg_dev_merge/dev_demo_no_logic_fixed
73+
als genai-utils --import-genai --using=../wg_demo_no_logic_fixed
74+
```
75+
76+
That will leave things in this state:
77+
78+
![genai-process](images/sample-ai/genai/genai-process-merged-models.png)
79+
80+
The `import-genai` command creates the `docs/import` directory and the following files, as shown above:
81+
82+
* `request.json` is sent to ChatGPT. It contains both models, and a command to merge them
83+
* `response.json` is the merged model. It should reflect the attributes from both sides, as shown
84+
* The response is translated to `system/genai/examples/genai_demo/wg_dev_merge/dev_demo_no_logic_fixed/docs/import/create_db_models.py`
85+
* The system creates `docs/import/create_db_models.py/create_db_models.sqlite` by executing the file above.
86+
* The system then uses this to update the dev project:
87+
* update the dev `database/db.sqlite` and
88+
* Runs `--rebuild-from-database`. This updates the model, the api, etc from the new database.
89+
* **It's good practice to verify these**. Make sure all the attributes from both sources are reflected in the updated **database** and **models** noted above.
90+
91+
<br/>
92+
93+
### Restart option for failure recovery
94+
95+
It may fail, requiring either a **re-run** or an `import-resume`:
96+
97+
* **Re-run** is indicated if the data model is missing attributes, incorrect or imcomplete.
98+
1. make sure to get initial `system/genai/examples/genai_demo/wg_dev_merge/dev_demo_no_logic_fixed/database/models.py` (eg, update from models_for_resume.py)
99+
2. delete or rename the `docs/import` directory.
100+
101+
* `import-resume` can be used if you can repair the file below, e.g., a minor syntax error.
102+
1. fix `system/genai/examples/genai_demo/wg_dev_merge/dev_demo_no_logic_fixed/docs/import/create_db_models.py`
103+
* Note: you can run this standalone with your IDE to verify it. It should create `create_db_models.sqlite` in your `docs/import` directory.
104+
2. make sure to get initial system/genai/examples/genai_demo/wg_dev_merge/dev_demo_no_logic_fixed/database/models.py (eg, update from models_for_resume.py)
105+
106+
```bash
107+
cd system/genai/examples/genai_demo/wg_dev_merge/dev_demo_no_logic_fixed
108+
als genai-utils --import-genai --using=../wg_demo_no_logic_fixed --import-resume
109+
```
110+
111+
<br/>
112+
113+
# Appendices
114+
115+
<br/>
116+
117+
## Ground Rules
118+
* No Dev_Team -> WG_Team integration (just deploy Dev_Team version, and use)
119+
* Dev team code cannot be integrated into WG - dependencies, libs, integration, ...
120+
* WG_Team - serial dev (as now)
121+
* WG_Team logic files are separate from Dev_Team (eg, using logic/discovery)
122+
* sqlite only, for now (presume upgrade to 'some other db' is doable later)
123+
* Tyler, what were the issues you mentioned in sqlite that forced you to use PG?
124+
* All Dev_Team and logic generations are finished before merge-G
125+
126+
<br/>
127+

‎docs/IDE-Rebuild-Test-Data.md

Lines changed: 10 additions & 6 deletions
Original file line numberDiff line numberDiff line change
@@ -1,24 +1,28 @@
11
!!! pied-piper ":bulb: TL;DR - Rebuild Test Data"
22

3-
Proper rule operation requires existing data be correct - see [Data ModelD esign](Data-Model-Design){:target="_blank" rel="noopener"}.
3+
Proper rule operation requires existing data be correct - see [Data Model Design](Data-Model-Design.md#Initialize Stored Derivations){:target="_blank" rel="noopener"}
44

5-
ChatGPT sometimes fails to properly compute the tests data. You can rebuild your test data to match the derivation rules by using `genai-utils --rebuild-test-data`.
5+
ChatGPT sometimes fails to properly compute the test data. You can rebuild your test data to match the derivation rules by using `genai-utils --rebuild-test-data`.
66

77

88
## Overview
99
Basic operation:
1010

11-
1. Builds database/test_data/test_data.py from docs/response.json
12-
2. Runs it to create database/test_data/db.sqlite
13-
3. Copies database/test_data/db.sqlite to database/db.sqlite
11+
1. Builds `database/test_data/test_data.py` from `docs/response.json`
12+
2. Runs it to create `database/test_data/db.sqlite`
13+
3. Copies `database/test_data/db.sqlite` to `database/db.sqlite`
1414

1515
&nbsp;
1616

1717
## Example
1818

1919
Fixes project issues by rebuilding the database to conform to the derivation rules:
2020

21-
1. Create genai_demo: `als genai --using=system/genai/examples/genai_demo/genai_demo.prompt --project-name=genai_demo`
21+
1. Create genai_demo:
22+
```
23+
als genai --using=system/genai/examples/genai_demo/genai_demo.prompt --project-name=genai_demo
24+
```
25+
2226
2. Rebuild:
2327
```
2428
cd genai_demo

0 commit comments

Comments
 (0)