[SYNC] Merge upstream NousResearch/hermes-agent — 499 commits #201

Merged
Rockachopa merged 500 commits from upstream-sync into main 2026-04-07 14:03:16 +00:00
6 changed files with 601 additions and 0 deletions
Showing only changes of commit 447ec076a4 - Show all commits

View File

@@ -120,3 +120,138 @@ self.play(old_content.animate.set_opacity(0.3), FadeIn(new_content))
self.play(FadeOut(Group(*self.mobjects)), run_time=0.5)
self.wait(0.3)
```
## Reactive Mobjects: always_redraw()
Rebuild a mobject from scratch every frame — essential when its geometry depends on other animated objects:
```python
# Brace that follows a resizing square
brace = always_redraw(Brace, square, UP)
self.add(brace)
self.play(square.animate.scale(2)) # brace auto-adjusts
# Horizontal line that tracks a moving dot
h_line = always_redraw(lambda: axes.get_h_line(dot.get_left()))
# Label that always stays next to another mobject
label = always_redraw(lambda: Text("here", font_size=20).next_to(dot, UP, buff=0.2))
```
Note: `always_redraw` recreates the mobject every frame. For simple property tracking, use `add_updater` instead (cheaper):
```python
label.add_updater(lambda m: m.next_to(dot, UP))
```
## TracedPath — Trajectory Tracing
Draw the path a point has traveled:
```python
dot = Dot(color=YELLOW)
path = TracedPath(dot.get_center, stroke_color=YELLOW, stroke_width=2)
self.add(dot, path)
self.play(dot.animate.shift(RIGHT * 3 + UP * 2), run_time=2)
# path shows the trail the dot left behind
# Fading trail (dissipates over time):
path = TracedPath(dot.get_center, dissipating_time=0.5, stroke_opacity=[0, 1])
```
Use cases: gradient descent paths, planetary orbits, function tracing, particle trajectories.
## FadeTransform — Smoother Cross-Fades
`Transform` morphs shapes through ugly intermediate warping. `FadeTransform` cross-fades with position matching — use it when source and target look different:
```python
# UGLY: Transform warps circle into square through a blob
self.play(Transform(circle, square))
# SMOOTH: FadeTransform cross-fades cleanly
self.play(FadeTransform(circle, square))
# FadeTransformPieces: per-submobject FadeTransform
self.play(FadeTransformPieces(group1, group2))
# TransformFromCopy: animate a COPY while keeping the original visible
self.play(TransformFromCopy(source, target))
# source stays on screen, a copy morphs into target
```
**Recommendation:** Use `FadeTransform` as default for dissimilar shapes. Use `Transform`/`ReplacementTransform` only for similar shapes (circle→ellipse, equation→equation).
## ApplyMatrix — Linear Transformation Visualization
Animate a matrix transformation on mobjects:
```python
# Apply a 2x2 matrix to a grid
matrix = [[2, 1], [1, 1]]
self.play(ApplyMatrix(matrix, number_plane), run_time=2)
# Also works on individual mobjects
self.play(ApplyMatrix([[0, -1], [1, 0]], square)) # 90-degree rotation
```
Pairs with `LinearTransformationScene` — see `camera-and-3d.md`.
## squish_rate_func — Time-Window Staggering
Compress any rate function into a time window within an animation. Enables overlapping stagger without `LaggedStart`:
```python
self.play(
FadeIn(a, rate_func=squish_rate_func(smooth, 0, 0.5)), # 0% to 50%
FadeIn(b, rate_func=squish_rate_func(smooth, 0.25, 0.75)), # 25% to 75%
FadeIn(c, rate_func=squish_rate_func(smooth, 0.5, 1.0)), # 50% to 100%
run_time=2
)
```
More precise than `LaggedStart` when you need exact overlap control.
## Additional Rate Functions
```python
from manim import (
smooth, linear, rush_into, rush_from,
there_and_back, there_and_back_with_pause,
running_start, double_smooth, wiggle,
lingering, exponential_decay, not_quite_there,
squish_rate_func
)
# running_start: pulls back before going forward (anticipation)
self.play(FadeIn(mob, rate_func=running_start))
# there_and_back_with_pause: goes there, holds, comes back
self.play(mob.animate.shift(UP), rate_func=there_and_back_with_pause)
# not_quite_there: stops at a fraction of the full animation
self.play(FadeIn(mob, rate_func=not_quite_there(0.7)))
```
## ShowIncreasingSubsets / ShowSubmobjectsOneByOne
Reveal group members progressively — ideal for algorithm visualization:
```python
# Reveal array elements one at a time
array = Group(*[Square() for _ in range(8)]).arrange(RIGHT)
self.play(ShowIncreasingSubsets(array), run_time=3)
# Show submobjects with staggered appearance
self.play(ShowSubmobjectsOneByOne(code_lines), run_time=4)
```
## ShowPassingFlash
A flash of light travels along a path:
```python
# Flash traveling along a curve
self.play(ShowPassingFlash(curve.copy().set_color(YELLOW), time_width=0.3))
# Great for: data flow, electrical signals, network traffic
```

View File

@@ -74,3 +74,62 @@ helix = ParametricFunction(
- Surfaces, vector fields, spatial geometry, 3D transforms
## When NOT to Use 3D
- 2D concepts, text-heavy scenes, flat data (bar charts, time series)
## ZoomedScene — Inset Zoom
Show a magnified inset of a detail while keeping the full view visible:
```python
class ZoomExample(ZoomedScene):
def __init__(self, **kwargs):
super().__init__(
zoom_factor=0.3, # how much of the scene the zoom box covers
zoomed_display_height=3, # size of the inset
zoomed_display_width=3,
zoomed_camera_frame_starting_position=ORIGIN,
**kwargs
)
def construct(self):
self.camera.background_color = BG
# ... create your scene content ...
# Activate the zoom
self.activate_zooming()
# Move the zoom frame to a point of interest
self.play(self.zoomed_camera.frame.animate.move_to(detail_point))
self.wait(2)
# Deactivate
self.play(self.get_zoomed_display_pop_out_animation(), rate_func=lambda t: smooth(1-t))
```
Use cases: zooming into a specific term in an equation, showing fine detail in a diagram, magnifying a region of a plot.
## LinearTransformationScene — Linear Algebra
Pre-built scene with basis vectors and grid for visualizing matrix transformations:
```python
class LinearTransformExample(LinearTransformationScene):
def __init__(self, **kwargs):
super().__init__(
show_coordinates=True,
show_basis_vectors=True,
**kwargs
)
def construct(self):
matrix = [[2, 1], [1, 1]]
# Add a vector before applying the transform
vector = self.get_vector([1, 2], color=YELLOW)
self.add_vector(vector)
# Apply the transformation — grid, basis vectors, and your vector all transform
self.apply_matrix(matrix)
self.wait(2)
```
This produces the signature 3Blue1Brown "Essence of Linear Algebra" look — grid lines deforming, basis vectors stretching, determinant visualized through area change.

View File

@@ -78,3 +78,88 @@ class DerivationScene(Scene):
s2.next_to(s1, DOWN, buff=0.8)
self.play(s1.animate.set_opacity(0.4), TransformMatchingTex(s1.copy(), s2))
```
## substrings_to_isolate for Complex Equations
For dense equations where manually splitting into parts is impractical, use `substrings_to_isolate` to tell Manim which substrings to track as individual elements:
```python
# Without isolation — the whole expression is one blob
lagrangian = MathTex(
r"\mathcal{L} = \bar{\psi}(i \gamma^\mu D_\mu - m)\psi - \tfrac{1}{4}F_{\mu\nu}F^{\mu\nu}"
)
# With isolation — each named substring is a separate submobject
lagrangian = MathTex(
r"\mathcal{L} = \bar{\psi}(i \gamma^\mu D_\mu - m)\psi - \tfrac{1}{4}F_{\mu\nu}F^{\mu\nu}",
substrings_to_isolate=[r"\psi", r"D_\mu", r"\gamma^\mu", r"F_{\mu\nu}"]
)
# Now you can color individual terms
lagrangian.set_color_by_tex(r"\psi", BLUE)
lagrangian.set_color_by_tex(r"F_{\mu\nu}", YELLOW)
```
Essential for `TransformMatchingTex` on complex equations — without isolation, matching fails on dense expressions.
## Multi-Line Complex Equations
For equations with multiple related lines, pass each line as a separate argument:
```python
maxwell = MathTex(
r"\nabla \cdot \mathbf{E} = \frac{\rho}{\epsilon_0}",
r"\nabla \times \mathbf{B} = \mu_0\mathbf{J} + \mu_0\epsilon_0\frac{\partial \mathbf{E}}{\partial t}"
).arrange(DOWN)
# Each line is a separate submobject — animate independently
self.play(Write(maxwell[0]))
self.wait(1)
self.play(Write(maxwell[1]))
```
## TransformMatchingTex with key_map
Map specific substrings between source and target equations during transformation:
```python
eq1 = MathTex(r"A^2 + B^2 = C^2")
eq2 = MathTex(r"A^2 = C^2 - B^2")
self.play(TransformMatchingTex(
eq1, eq2,
key_map={"+": "-"}, # map "+" in source to "-" in target
path_arc=PI / 2, # arc the pieces into position
))
```
## set_color_by_tex — Color by Substring
```python
eq = MathTex(r"E = mc^2")
eq.set_color_by_tex("E", BLUE)
eq.set_color_by_tex("m", RED)
eq.set_color_by_tex("c", GREEN)
```
## TransformMatchingTex with matched_keys
When matching substrings are ambiguous, specify which to align explicitly:
```python
kw = dict(font_size=72, t2c={"A": BLUE, "B": TEAL, "C": GREEN})
lines = [
MathTex(r"A^2 + B^2 = C^2", **kw),
MathTex(r"A^2 = C^2 - B^2", **kw),
MathTex(r"A^2 = (C + B)(C - B)", **kw),
MathTex(r"A = \sqrt{(C + B)(C - B)}", **kw),
]
self.play(TransformMatchingTex(
lines[0].copy(), lines[1],
matched_keys=["A^2", "B^2", "C^2"], # explicitly match these
key_map={"+": "-"}, # map + to -
path_arc=PI / 2, # arc pieces into position
))
```
Without `matched_keys`, the animation matches the longest common substrings, which can produce unexpected results on complex equations (e.g., "^2 = C^2" matching across terms).

View File

@@ -89,3 +89,75 @@ arrow = Arrow(before.get_right(), after.get_left(), color=YELLOW)
label = Text("+167%", font_size=36, color=YELLOW).next_to(arrow, UP)
self.play(GrowArrow(arrow), Write(label))
```
## Graph / DiGraph — Graph Theory Visualization
Built-in graph mobjects with automatic layout:
```python
# Undirected graph
g = Graph(
vertices=[1, 2, 3, 4, 5],
edges=[(1, 2), (2, 3), (3, 4), (4, 5), (5, 1), (1, 3)],
layout="spring", # or "circular", "kamada_kawai", "planar", "tree"
labels=True,
vertex_config={"fill_color": PRIMARY},
edge_config={"stroke_color": SUBTLE},
)
self.play(Create(g))
# Directed graph
dg = DiGraph(
vertices=["A", "B", "C"],
edges=[("A", "B"), ("B", "C"), ("C", "A")],
layout="circular",
labels=True,
edge_config={("A", "B"): {"stroke_color": RED}},
)
# Add/remove vertices and edges dynamically
self.play(g.animate.add_vertices(6, positions={6: RIGHT * 2}))
self.play(g.animate.add_edges((1, 6)))
self.play(g.animate.remove_vertices(3))
```
Layout algorithms: `"spring"`, `"circular"`, `"kamada_kawai"`, `"planar"`, `"spectral"`, `"tree"` (for rooted trees, specify `root=`).
## ArrowVectorField / StreamLines — Vector Fields
```python
# Arrow field: arrows showing direction at each point
field = ArrowVectorField(
lambda pos: np.array([-pos[1], pos[0], 0]), # rotation field
x_range=[-3, 3], y_range=[-3, 3],
colors=[BLUE, GREEN, YELLOW, RED]
)
self.play(Create(field))
# StreamLines: flowing particle traces through the field
stream = StreamLines(
lambda pos: np.array([-pos[1], pos[0], 0]),
stroke_width=2, max_anchors_per_line=30
)
self.add(stream)
stream.start_animation(warm_up=True, flow_speed=1.5)
self.wait(3)
stream.end_animation()
```
Use cases: electromagnetic fields, fluid flow, gradient fields, ODE phase portraits.
## ComplexPlane / PolarPlane
```python
# Complex plane with Re/Im labels
cplane = ComplexPlane().add_coordinates()
dot = Dot(cplane.n2p(2 + 1j), color=YELLOW)
label = Text("2+i", font_size=20).next_to(dot, UR, buff=0.1)
# Apply complex function to the plane
self.play(cplane.animate.apply_complex_function(lambda z: z**2), run_time=3)
# Polar plane
polar = PolarPlane(radius_max=3).add_coordinates()
```

View File

@@ -104,3 +104,161 @@ class NetworkNode(Group):
Directions: `UP, DOWN, LEFT, RIGHT, ORIGIN, UL, UR, DL, DR`
Colors: `RED, BLUE, GREEN, YELLOW, WHITE, GRAY, ORANGE, PINK, PURPLE, TEAL, GOLD`
Frame: `config.frame_width = 14.222, config.frame_height = 8.0`
## SVGMobject — Import SVG Files
```python
logo = SVGMobject("path/to/logo.svg")
logo.set_color(WHITE).scale(0.5).to_corner(UR)
self.play(FadeIn(logo))
# SVG submobjects are individually animatable
for part in logo.submobjects:
self.play(part.animate.set_color(random_color()))
```
## ImageMobject — Display Images
```python
img = ImageMobject("screenshot.png")
img.set_height(3).to_edge(RIGHT)
self.play(FadeIn(img))
```
Note: images cannot be animated with `.animate` (they're raster, not vector). Use `FadeIn`/`FadeOut` and `shift`/`scale` only.
## Variable — Auto-Updating Display
```python
var = Variable(0, Text("x"), num_decimal_places=2)
var.move_to(ORIGIN)
self.add(var)
# Animate the value
self.play(var.tracker.animate.set_value(5), run_time=2)
# Display auto-updates: "x = 5.00"
```
Cleaner than manual `DecimalNumber` + `add_updater` for simple labeled-value displays.
## BulletedList
```python
bullets = BulletedList(
"First key point",
"Second important fact",
"Third conclusion",
font_size=28
)
bullets.to_edge(LEFT, buff=1.0)
self.play(Write(bullets))
# Highlight individual items
self.play(bullets[1].animate.set_color(YELLOW))
```
## DashedLine and Angle Markers
```python
# Dashed line (asymptotes, construction lines)
dashed = DashedLine(LEFT * 3, RIGHT * 3, color=SUBTLE, dash_length=0.15)
# Angle marker between two lines
line1 = Line(ORIGIN, RIGHT * 2)
line2 = Line(ORIGIN, UP * 2 + RIGHT)
angle = Angle(line1, line2, radius=0.5, color=YELLOW)
angle_label = angle.get_value() # returns the angle in radians
# Right angle marker
right_angle = RightAngle(line1, Line(ORIGIN, UP * 2), length=0.3, color=WHITE)
```
## Boolean Operations (CSG)
Combine, subtract, or intersect 2D shapes:
```python
circle = Circle(radius=1.5, color=BLUE, fill_opacity=0.5).shift(LEFT * 0.5)
square = Square(side_length=2, color=RED, fill_opacity=0.5).shift(RIGHT * 0.5)
# Union, Intersection, Difference, Exclusion
union = Union(circle, square, color=GREEN, fill_opacity=0.5)
intersect = Intersection(circle, square, color=YELLOW, fill_opacity=0.5)
diff = Difference(circle, square, color=PURPLE, fill_opacity=0.5)
exclude = Exclusion(circle, square, color=ORANGE, fill_opacity=0.5)
```
Use cases: Venn diagrams, set theory, geometric proofs, area calculations.
## LabeledArrow / LabeledLine
```python
# Arrow with built-in label (auto-positioned)
arr = LabeledArrow(Text("force", font_size=18), start=LEFT, end=RIGHT, color=RED)
# Line with label
line = LabeledLine(Text("d = 5m", font_size=18), start=LEFT * 2, end=RIGHT * 2)
```
Auto-handles label positioning — cleaner than manual `Arrow` + `Text().next_to()`.
## Text Color/Font/Style Per-Substring (t2c, t2f, t2s, t2w)
```python
# Color specific words (t2c = text-to-color)
text = Text(
"Gradient descent minimizes the loss function",
t2c={"Gradient descent": BLUE, "loss function": RED}
)
# Different fonts per word (t2f = text-to-font)
text = Text(
"Use Menlo for code and Inter for prose",
t2f={"Menlo": "Menlo", "Inter": "Inter"}
)
# Italic/slant per word (t2s = text-to-slant)
text = Text("Normal and italic text", t2s={"italic": ITALIC})
# Bold per word (t2w = text-to-weight)
text = Text("Normal and bold text", t2w={"bold": BOLD})
```
These are much cleaner than creating separate Text objects and grouping them.
## Backstroke for Readability Over Backgrounds
When text overlaps other content (graphs, diagrams, images), add a dark stroke behind it:
```python
# CE syntax:
label.set_stroke(BLACK, width=5, background=True)
# Apply to a group
for mob in labels:
mob.set_stroke(BLACK, width=4, background=True)
```
This is how 3Blue1Brown keeps text readable over complex backgrounds without using BackgroundRectangle.
## Complex Function Transforms
Apply complex functions to entire mobjects — transforms the plane:
```python
c_grid = ComplexPlane()
moving_grid = c_grid.copy()
moving_grid.prepare_for_nonlinear_transform() # adds more sample points for smooth deformation
self.play(
moving_grid.animate.apply_complex_function(lambda z: z**2),
run_time=5,
)
# Also works with R3->R3 functions:
self.play(grid.animate.apply_function(
lambda p: [p[0] + 0.5 * math.sin(p[1]), p[1] + 0.5 * math.sin(p[0]), p[2]]
), run_time=5)
```
**Critical:** Call `prepare_for_nonlinear_transform()` before applying nonlinear functions — without it, the grid has too few sample points and the deformation looks jagged.

View File

@@ -91,3 +91,95 @@ manim -ql --resolution 1080,1080 script.py Scene # 1:1 square
5. Review stitched output
6. Production render at `-qh`
7. Re-stitch + add audio
## manim.cfg — Project Configuration
Create `manim.cfg` in the project directory for per-project defaults:
```ini
[CLI]
quality = low_quality
preview = True
media_dir = ./media
[renderer]
background_color = #0D1117
[tex]
tex_template_file = custom_template.tex
```
This eliminates repetitive CLI flags and `self.camera.background_color` in every scene.
## Sections — Chapter Markers
Mark sections within a scene for organized output:
```python
class LongVideo(Scene):
def construct(self):
self.next_section("Introduction")
# ... intro content ...
self.next_section("Main Concept")
# ... main content ...
self.next_section("Conclusion")
# ... closing ...
```
Render individual sections: `manim --save_sections script.py LongVideo`
This outputs separate video files per section — useful for long videos where you want to re-render only one part.
## manim-voiceover Plugin (Recommended for Narrated Videos)
The official `manim-voiceover` plugin integrates TTS directly into scene code, auto-syncing animation duration to voiceover length. This is significantly cleaner than the manual ffmpeg muxing approach above.
### Installation
```bash
pip install "manim-voiceover[elevenlabs]"
# Or for free/local TTS:
pip install "manim-voiceover[gtts]" # Google TTS (free, lower quality)
pip install "manim-voiceover[azure]" # Azure Cognitive Services
```
### Usage
```python
from manim import *
from manim_voiceover import VoiceoverScene
from manim_voiceover.services.elevenlabs import ElevenLabsService
class NarratedScene(VoiceoverScene):
def construct(self):
self.set_speech_service(ElevenLabsService(
voice_name="Alice",
model_id="eleven_multilingual_v2"
))
# Voiceover auto-controls scene duration
with self.voiceover(text="Here is a circle being drawn.") as tracker:
self.play(Create(Circle()), run_time=tracker.duration)
with self.voiceover(text="Now let's transform it into a square.") as tracker:
self.play(Transform(circle, Square()), run_time=tracker.duration)
```
### Key Features
- `tracker.duration` — total voiceover duration in seconds
- `tracker.time_until_bookmark("mark1")` — sync specific animations to specific words
- Auto-generates subtitle `.srt` files
- Caches audio locally — re-renders don't re-generate TTS
- Works with: ElevenLabs, Azure, Google TTS, pyttsx3 (offline), and custom services
### Bookmarks for Precise Sync
```python
with self.voiceover(text='This is a <bookmark mark="circle"/>circle.') as tracker:
self.wait_until_bookmark("circle")
self.play(Create(Circle()), run_time=tracker.time_until_bookmark("circle", limit=1))
```
This is the recommended approach for any video with narration. The manual ffmpeg muxing workflow above is still useful for adding background music or post-production audio mixing.