Render a HTML Report for GlySmith Results

Generate a self-contained HTML report for a glysmith_result object. The report is rendered via rmarkdown::render() using an internal R Markdown template. If use_ai is TRUE, the report text will be polished and organized into sections using LLM (deepseek-chat), and plots will be described with a multimodal model (deepseek-vl-chat). To use this feature, you have to provide an API key and set it in the environment variable DEEPSEEK_API_KEY by running Sys.setenv(DEEPSEEK_API_KEY = "your_api_key"). You can apply the API key on https://platform.deepseek.com.

Usage

polish_report(
  x,
  output_file,
  title = "GlySmith report",
  open = interactive(),
  use_ai = FALSE
)

Arguments

x

A glysmith_result object.

output_file

Path to the output HTML file.

title

Report title.

open

Whether to open the report in a browser after rendering.

use_ai

Whether to polish the report text, organize sections, and generate plot descriptions using AI (deepseek-chat and deepseek-vision). Default is FALSE.

Value

The normalized path to the generated HTML file.

Examples

library(glyexp)
exp <- real_experiment2
result <- forge_analysis(exp)
#> ℹ Identification overview
#> ✔ Identification overview [129ms]
#> 
#> ℹ Preprocessing
#> 
#> ℹ Preprocessing

#> ── Removing variables with too many missing values ──
#> ℹ Preprocessing

#> 
#> ℹ Preprocessing

#> ℹ No QC samples found. Using all samples.
#> ℹ Preprocessing

#> ℹ Applying preset "discovery"...
#> ℹ Preprocessing

#> ℹ Total removed: 10 (14.93%) variables.
#> ℹ Preprocessing

#> ✔ Variable removal completed.
#> ℹ Preprocessing

#> 
#> ℹ Preprocessing

#> ── Normalizing data ──
#> ℹ Preprocessing

#> 
#> ℹ Preprocessing

#> ℹ No QC samples found. Using default normalization method based on experiment type.
#> ℹ Preprocessing

#> ℹ Experiment type is "glycomics". Using `normalize_median_quotient()` + `normalize_total_area()`.
#> ℹ Preprocessing

#> ✔ Normalization completed.
#> ℹ Preprocessing

#> 
#> ℹ Preprocessing

#> ── Normalizing data (Total Area) ──
#> ℹ Preprocessing

#> 
#> ℹ Preprocessing

#> ✔ Total area normalization completed.
#> ℹ Preprocessing

#> 
#> ℹ Preprocessing

#> ── Imputing missing values ──
#> ℹ Preprocessing

#> 
#> ℹ Preprocessing

#> ℹ No QC samples found. Using default imputation method based on sample size.
#> ℹ Preprocessing

#> ℹ Sample size > 100, using `impute_miss_forest()`.
#> ℹ Preprocessing

#> ✔ Imputation completed.
#> ℹ Preprocessing

#> 
#> ℹ Preprocessing

#> ── Correcting batch effects ──
#> ℹ Preprocessing

#> 
#> ℹ Preprocessing

#> ℹ Batch column  not found in sample_info. Skipping batch correction.
#> ℹ Preprocessing

#> ✔ Batch correction completed.
#> ℹ Preprocessing

#> ✔ Preprocessing [5.6s]
#> 
#> ℹ QC (post-preprocessing)
#> ✔ QC (post-preprocessing) [126ms]
#> 
#> ℹ Principal component analysis
#> ✔ Principal component analysis [300ms]
#> 
#> ℹ Differential expression analysis (limma)
#> ℹ Number of groups: 4
#> ℹ Differential expression analysis (limma)

#> ℹ Groups: "H", "M", "Y", and "C"
#> ℹ Differential expression analysis (limma)

#> ℹ Pairwise comparisons will be performed, with levels coming first as reference groups.
#> ℹ Differential expression analysis (limma)

#> ✔ Differential expression analysis (limma) [58ms]
#> 
#> ℹ Volcano plot
#> ✔ Volcano plot [536ms]
#> 
#> ℹ Heatmap of significant variables
#> ✔ Heatmap of significant variables [66ms]
#> 
#> ℹ Skipping `step_sig_enrich_go()` because input is not a glycoproteomics experiment and input has more than 2 groups.
#> ℹ Skipping `step_sig_enrich_kegg()` because input is not a glycoproteomics experiment and input has more than 2 groups.
#> ℹ Skipping `step_sig_enrich_reactome()` because input is not a glycoproteomics experiment and input has more than 2 groups.
#> ℹ Derived trait calculation
#> ✔ Derived trait calculation [2.2s]
#> 
#> ℹ Differential trait analysis (limma)
#> ℹ Number of groups: 4
#> ℹ Differential trait analysis (limma)

#> ℹ Groups: "H", "M", "Y", and "C"
#> ℹ Differential trait analysis (limma)

#> ℹ Pairwise comparisons will be performed, with levels coming first as reference groups.
#> ℹ Differential trait analysis (limma)

#> ✔ Differential trait analysis (limma) [56ms]
#> 
#> ℹ Heatmap of significant traits
#> ✔ Heatmap of significant traits [74ms]
#> 
polish_report(result, tempfile(fileext = ".html"), open = FALSE)
#> [1] "/tmp/RtmpBgHsNs/file1c902434a622.html"