Recently, a lot of open source e-learning platforms have been offered for free in the Internet. We thus incorporate the intelligent diagnosis and assessment tool into an open software e-learning platform developed for programming language courses, wherein the proposed learning diagnosis assessment tools based on text mining and machine learning techniques are employed to alleviate the loading of the teachers. Experiments were conducted in two introductory-undergraduate programming courses to examine the effectiveness of the proposed diagnosis and assessment tools. The learners' work including the source code and comments were processed by the proposed text mining and machine learning techniques. This system also provides immediate feedback and high-quality evaluation results to guide the learners with poor performance. Our experimental results reveal that the proposed work can effectively assist the low-ability learners. (Contains 6 tables and 13 figures.)