@imstudium/cli 0.1.1 → 0.1.3
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
package/package.json
CHANGED
|
@@ -1,6 +1,6 @@
|
|
|
1
1
|
{
|
|
2
2
|
"name": "@imstudium/cli",
|
|
3
|
-
"version": "0.1.
|
|
3
|
+
"version": "0.1.3",
|
|
4
4
|
"description": "ImStudium CLI — study workspace from Stud.IP",
|
|
5
5
|
"license": "MIT",
|
|
6
6
|
"type": "module",
|
|
@@ -26,7 +26,7 @@
|
|
|
26
26
|
"opencode"
|
|
27
27
|
],
|
|
28
28
|
"bin": {
|
|
29
|
-
"imstudium": "
|
|
29
|
+
"imstudium": "dist/cli.js"
|
|
30
30
|
},
|
|
31
31
|
"files": [
|
|
32
32
|
"dist",
|
|
@@ -40,7 +40,7 @@
|
|
|
40
40
|
"prepublishOnly": "bun run build && bun run typecheck"
|
|
41
41
|
},
|
|
42
42
|
"dependencies": {
|
|
43
|
-
"@imstudium/sdk": "
|
|
43
|
+
"@imstudium/sdk": "^0.1.3",
|
|
44
44
|
"commander": "^13.1.0",
|
|
45
45
|
"playwright": "^1.52.0"
|
|
46
46
|
},
|
|
@@ -48,7 +48,9 @@ imstudium files tree --course "Analysis"
|
|
|
48
48
|
|
|
49
49
|
## Reading slides
|
|
50
50
|
|
|
51
|
-
After `extract`, each file has `extracted/<file>/content.md` and `pages/*.png`.
|
|
51
|
+
After `extract`, each file has `extracted/<file>/content.md` and `pages/*.png`.
|
|
52
|
+
|
|
53
|
+
**Limits:** text is poppler/OCR — not LaTeX markdown. OCR fails on handwritten math. Use PNGs for formulas and diagrams (see `imstudium-visual` skill).
|
|
52
54
|
|
|
53
55
|
## Auth
|
|
54
56
|
|
|
@@ -56,4 +56,10 @@ Prefer local tools before `imstudium_sync` (network + auth).
|
|
|
56
56
|
|
|
57
57
|
## Visual comprehension (important)
|
|
58
58
|
|
|
59
|
-
`studium_read_material` returns `content` (text) and `pageImages` (PNG paths).
|
|
59
|
+
`studium_read_material` returns `content` (text) and `pageImages` (PNG paths).
|
|
60
|
+
|
|
61
|
+
**No LaTeX in extracted text** — poppler/Tesseract only. `content.md` and corpus hits are for discovery; formulas are often wrong or missing in text.
|
|
62
|
+
|
|
63
|
+
Slides carry diagrams, plots, and formulas that text misses — **read the `pageImages` files** with your image-reading tool before explaining slide content. Use the `page` argument to target one slide.
|
|
64
|
+
|
|
65
|
+
`studium_overview` → `searchGuidance.extractionLimits` for the full list.
|
|
@@ -14,6 +14,21 @@ allowed-tools:
|
|
|
14
14
|
|
|
15
15
|
Slides are **visual**. `content.md` text misses diagrams, plots, formulas, tables, and layout.
|
|
16
16
|
|
|
17
|
+
## Extraction limits (know this)
|
|
18
|
+
|
|
19
|
+
ImStudium does **not** produce LaTeX `$...$` in markdown. Local tools only:
|
|
20
|
+
|
|
21
|
+
| Source | What you get in `content.md` |
|
|
22
|
+
|--------|------------------------------|
|
|
23
|
+
| Typed PDF | poppler Unicode text — math symbols at best, not LaTeX |
|
|
24
|
+
| Scanned / handwritten | Tesseract OCR — **unreliable for formulas** (`_(OCR)_` prefix) |
|
|
25
|
+
| PPTX (XML fallback) | Slide text only — Office Math omitted |
|
|
26
|
+
| `.tex` source files | LaTeX passthrough as-is |
|
|
27
|
+
|
|
28
|
+
**corpus.jsonl** is for finding pages by keyword — not for answering math from OCR text alone.
|
|
29
|
+
|
|
30
|
+
For any formula, diagram, plot, or table: **read `pageImages` or open `viewerFileUrl`** before explaining.
|
|
31
|
+
|
|
17
32
|
## Rule (all agents)
|
|
18
33
|
|
|
19
34
|
**Never summarize slide content from text alone.** Before answering:
|