This paper introduces a new method for the digital and computational analysis of historical parliamentary proceedings. The paper presents an XML-TEI model specifically designed for encoding historical parliamentary documents; this model is exemplified through the analysis of parliamentary debates from the French Chamber of Deputies (1889-1893). The first part of the paper discusses the motivations behind the model's development. The second part outlines the methodological choices in constructing the model and the need for schema adaptation. We subsequently detail our method for automatic encoding of such large corpora. Finally, we propose an approach to annotate parliamentary debates using natural language processing analyses, focusing on topic modelling.