Merge pull request #177 from renecotyfanboy/v0.1.0

V0.1.0

docs/examples/background.md

@@ -12,8 +12,8 @@ approach is equivalent to subtract the background to the observed spectrum when
 ``` python
 from jaxspec.model.background import SubtractedBackground
 
-forward = NUTSFitter(model, obs, background_model=SubtractedBackground())
-result = forward.fit(prior, num_chains=4, num_warmup=1000, num_samples=1000, mcmc_kwargs={"progress_bar": True})
+fitter = NUTSFitter(model, prior, obs, background_model=SubtractedBackground())
+result = fitter.fit(num_chains=4, num_warmup=1000, num_samples=1000, mcmc_kwargs={"progress_bar": True})
 
 result.plot_ppc()
 ```
@@ -28,8 +28,8 @@ it is to consider each background bin as a Poisson realisation of a counting pro
 ``` python
 from jaxspec.model.background import BackgroundWithError
 
-forward = NUTSFitter(model, obs, background_model=BackgroundWithError())
-result = forward.fit(prior, num_chains=4, num_warmup=1000, num_samples=1000, mcmc_kwargs={"progress_bar": True})
+fitter = NUTSFitter(model, prior, obs, background_model=BackgroundWithError())
+result = fitter.fit(num_chains=4, num_warmup=1000, num_samples=1000, mcmc_kwargs={"progress_bar": True})
 
 result.plot_ppc()
 ```
@@ -43,8 +43,8 @@ nodes will drive the flexibility of the Gaussian process, and it should always b
 ``` python
 from jaxspec.model.background import GaussianProcessBackground
 
-forward = NUTSFitter(model, obs, background_model=GaussianProcessBackground(e_min=0.3, e_max=8, n_nodes=20))
-result = forward.fit(prior, num_chains=4, num_warmup=1000, num_samples=1000, mcmc_kwargs={"progress_bar": True})
+forward = NUTSFitter(model, prior, obs, background_model=GaussianProcessBackground(e_min=0.3, e_max=8, n_nodes=20))
+result = forward.fit(num_chains=4, num_warmup=1000, num_samples=1000, mcmc_kwargs={"progress_bar": True})
 
 result.plot_ppc()
 ```

docs/examples/index.md

@@ -64,4 +64,11 @@
 
     Add a custom model to `jaxspec`
 
+-   [__Interface with other frameworks__](external_samplers.ipynb)
+
+    ---
+
+    Use `jaxspec` likelihood in other packages
+
+
 </div>

mkdocs.yml

@@ -16,6 +16,7 @@ nav:
     - Add a background : examples/background.md
     - Work with arviz : examples/result_analysis.ipynb
     - Build a custom model : examples/build_custom_model.ipynb
+    - Interface with other frameworks : examples/external_samplers.ipynb
   #- Models:
   #    - models/index.md
   #    - APEC: models/apec.md

pyproject.toml

@@ -1,10 +1,13 @@
 [tool.poetry]
 name = "jaxspec"
-version = "0.0.8"
+version = "0.1.0"
 description = "jaxspec is a bayesian spectral fitting library for X-ray astronomy."
 authors = ["sdupourque <sdupourque@irap.omp.eu>"]
 license = "MIT"
 readme = "README.md"
+homepage = "https://github.com/renecotyfanboy/jaxspec"
+documentation = "https://jaxspec.readthedocs.io/en/latest/"
+
 
 [tool.poetry.dependencies]
 python = ">=3.10,<3.12"
@@ -13,12 +16,12 @@ jaxlib = "^0.4.30"
 numpy = "<2.0.0"
 pandas = "^2.2.0"
 astropy = "^6.0.0"
-numpyro = "^0.15.0"
+numpyro = "^0.15.2"
 dm-haiku = ">=0.0.11,<0.0.13"
 networkx = "^3.1"
 matplotlib = "^3.8.0"
 arviz = ">=0.17.1,<0.20.0"
-chainconsumer = "^1.0.0"
+chainconsumer = "^1.1.2"
 simpleeval = "^0.9.13"
 cmasher = "^1.6.3"
 gpjax = "^0.8.0"
@@ -34,6 +37,7 @@ pyzmq = "<27"
 jaxns = "^2.5.1"
 pooch = "^1.8.2"
 interpax = "^0.3.3"
+watermark = "^2.4.3"
 
 
 [tool.poetry.group.docs.dependencies]
@@ -112,3 +116,6 @@ lines-between-types = 1
 [build-system]
 requires = ["poetry-core"]
 build-backend = "poetry.core.masonry.api"
+
+[tool.poetry.scripts]
+jaxspec-debug-info = "jaxspec.scripts.debug:debug_info"

src/jaxspec/analysis/results.py

@@ -615,6 +615,17 @@ def plot_ppc(
                     y_observed_bkg = (
                         obsconf.folded_background.data * u.photon / (denominator * ratio)
                     ).to(y_units)
+                    y_observed_bkg_low = (
+                        nbinom.ppf(percentile[0] / 100, obsconf.folded_background.data, 0.5)
+                        * u.photon
+                        / (denominator * ratio)
+                    ).to(y_units)
+                    y_observed_bkg_high = (
+                        nbinom.ppf(percentile[1] / 100, obsconf.folded_background.data, 0.5)
+                        * u.photon
+                        / (denominator * ratio)
+                    ).to(y_units)
+
                     legend_plots += _plot_binned_samples_with_error(
                         ax[0],
                         xbins.value,
@@ -626,6 +637,23 @@ def plot_ppc(
 
                     legend_labels.append("Model (bkg)")
 
+                    true_bkg_plot = ax[0].errorbar(
+                        np.sqrt(xbins.value[0] * xbins.value[1]),
+                        y_observed_bkg.value,
+                        xerr=np.abs(xbins.value - np.sqrt(xbins.value[0] * xbins.value[1])),
+                        yerr=[
+                            y_observed_bkg.value - y_observed_bkg_low.value,
+                            y_observed_bkg_high.value - y_observed_bkg.value,
+                        ],
+                        color="black",
+                        linestyle="none",
+                        alpha=0.3,
+                        capsize=2,
+                    )
+
+                    legend_plots.append((true_bkg_plot,))
+                    legend_labels.append("Observed (bkg)")
+
                 residual_samples = (obsconf.folded_counts.data - count) / np.diff(
                     np.percentile(count, percentile, axis=0), axis=0
                 )